pref_extract_more_inputs.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "pref_extract_more_inputs.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "d30Y6Tjy9j89",
        "colab_type": "text"
      },
      "source": [
        "Uses the reduction from the rl util to infer human preferences. **Run the export_activations_reductions.ipynb notebook before this one** to get the reduction saved in your drive."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zYEIvaw755Ou",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import joblib\n",
        "import os\n",
        "import itertools\n",
        "import random\n",
        "import time\n",
        "from sklearn.utils import shuffle\n",
        "import numpy as np\n",
        "import sys\n",
        "\n",
        "%tensorflow_version 2.x\n",
        "import tensorflow as tf"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "n9FqWdiv6drn",
        "colab_type": "code",
        "outputId": "76e2333b-408f-463f-90fc-48362a6d2383",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 122
        }
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Go to this URL in a browser: https://accounts.google.com/o/oauth2/auth?client_id=947318989803-6bn6qk8qdgf4n4g3pfee6491hc0brc4i.apps.googleusercontent.com&redirect_uri=urn%3aietf%3awg%3aoauth%3a2.0%3aoob&response_type=code&scope=email%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdocs.test%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive%20https%3a%2f%2fwww.googleapis.com%2fauth%2fdrive.photos.readonly%20https%3a%2f%2fwww.googleapis.com%2fauth%2fpeopleapi.readonly\n",
            "\n",
            "Enter your authorization code:\n",
            "··········\n",
            "Mounted at /content/drive\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "C8Xdv5Sv6rXA",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "file_path = '/content/drive/My Drive/pref_extract_files/'\n",
        "ys = joblib.load(os.path.join(file_path, 'ys.pkl'))\n",
        "observations = joblib.load(os.path.join(file_path, 'observations.pkl'))\n",
        "dense_activations = joblib.load(os.path.join(file_path, 'dense_activations.pkl'))\n",
        "conv_activations = joblib.load(os.path.join(file_path, 'conv_activations.pkl'))\n",
        "conv_reduced = joblib.load(os.path.join(file_path, 'reduced_train_attri_opt_act.pkl'))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_5i2FCnrFXFC",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def normalise(x):\n",
        "  x -= np.mean(x)\n",
        "  x /= np.std(x)\n",
        "  return x\n",
        "\n",
        "observations = normalise(observations)\n",
        "dense_activations = normalise(dense_activations)\n",
        "conv_activations = normalise(conv_activations)\n",
        "conv_reduced = normalise(conv_reduced)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "geB-dPGnrW_A",
        "colab_type": "code",
        "outputId": "3eac627a-fa52-40d0-9ccc-c537c63db1f3",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "print(dense_activations.shape)\n",
        "\n",
        "# Change for different training data\n",
        "img_like_for_x = conv_reduced\n",
        "# img_like_for_x = conv_activations\n",
        "print(img_like_for_x.shape)\n",
        "\n",
        "xs = []\n",
        "for img_like, dense_activation in zip(img_like_for_x, dense_activations):\n",
        "  xs.append([img_like, dense_activation])"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(4000, 64)\n",
            "(4000, 12, 14, 6)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dUuEUk_G65pi",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def get_val_auc(logs):\n",
        "      for key in logs:\n",
        "        if key.startswith('val_auc'):\n",
        "          return logs[key]\n",
        "\n",
        "class BestStats(tf.keras.callbacks.Callback):\n",
        "  \"\"\"A callback to keep track of the best val accuracy and auc seen so far.\"\"\"\n",
        "  def on_train_begin(self, logs):\n",
        "      self.bestMetric = -float('inf')\n",
        "      self.bestLogs = None\n",
        "      self.bestTrain = -float('inf')\n",
        "      self.num_epochs = 0\n",
        "\n",
        "  def on_epoch_end(self, epoch, logs):\n",
        "    self.num_epochs += 1\n",
        "    self.bestTrain = max(self.bestTrain, logs.get('accuracy'))\n",
        "\n",
        "    val_accuracy = logs.get('val_accuracy')\n",
        "    if val_accuracy == None:\n",
        "      return \n",
        "\n",
        "    val_auc = get_val_auc(logs)\n",
        "    \n",
        "    metric = (val_accuracy + val_auc) / 2.0\n",
        "\n",
        "    if metric > self.bestMetric:\n",
        "      self.bestMetric = metric\n",
        "      self.bestLogs = logs"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "E2cuwyUg7Sss",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def get_model(reg_amount, drop_rate, do_ave_pool, conv_layer_params, fc_layer_sizes, concat_top_layer, reduction_alg, n_components):\n",
        "  del reduction_alg, n_components\n",
        "  \"\"\"\n",
        "  Args:\n",
        "    conv_layer_params: Optional list of convolution layers parameters, where\n",
        "      each item is either a length-three tuple indicating\n",
        "      `(filters, kernel_size, stride)`.\n",
        "      \"\"\"\n",
        "\n",
        "  conv_model = tf.keras.models.Sequential()\n",
        "\n",
        "  for config in conv_layer_params:\n",
        "    (filters, kernel_size, strides) = config\n",
        "    conv_model.add(\n",
        "        tf.keras.layers.Conv2D(\n",
        "            filters=filters,\n",
        "            kernel_size=kernel_size,\n",
        "            strides=strides,\n",
        "            activation='relu',\n",
        "            ))\n",
        "  if do_ave_pool:\n",
        "    conv_model.add(tf.keras.layers.GlobalAveragePooling2D())\n",
        "  conv_model.add(tf.keras.layers.Flatten())\n",
        "\n",
        "  dense_model = tf.keras.models.Sequential(name='output')\n",
        "  for layer_size in fc_layer_sizes:\n",
        "    dense_model.add(tf.keras.layers.Dense(layer_size, activation='relu', kernel_regularizer=tf.keras.regularizers.l2(reg_amount)))\n",
        "    dense_model.add(tf.keras.layers.Dropout(drop_rate))\n",
        "  dense_model.add(tf.keras.layers.Dense(1, activation='sigmoid'))\n",
        "\n",
        "  img_input = tf.keras.layers.Input(shape=img_like_for_x.shape[1:], name='img_input')\n",
        "  dense_input = tf.keras.layers.Input(shape=dense_activations.shape[1:], name='dense_input')\n",
        "\n",
        "  if concat_top_layer:\n",
        "    final_img_input = img_input\n",
        "  else:\n",
        "    # Tile the dense input to be the same size as img_input, and concat them.\n",
        "    dense_input_expand = tf.keras.layers.Reshape([1, 1, dense_input.shape[-1]])(dense_input)\n",
        "    dense_input_tile = tf.keras.backend.tile(dense_input_expand, [1, img_input.shape[1], img_input.shape[2], 1])\n",
        "    final_img_input = tf.keras.layers.concatenate([img_input, dense_input_tile])\n",
        "\n",
        "  conv_output = conv_model(final_img_input)\n",
        "  \n",
        "  if concat_top_layer:\n",
        "    # Concatenate the dense_input to the output of the conv model.\n",
        "    final_dense_input = tf.keras.layers.concatenate([conv_output, dense_input])\n",
        "  else:\n",
        "    final_dense_input = conv_output\n",
        "    \n",
        "  output = dense_model(final_dense_input)\n",
        "\n",
        "  model = tf.keras.models.Model(inputs=[img_input, dense_input], outputs=[output])\n",
        "\n",
        "  model.compile(optimizer=tf.keras.optimizers.Adam(.01),\n",
        "                loss='binary_crossentropy',\n",
        "                metrics=['accuracy', tf.keras.metrics.AUC()],\n",
        "                )\n",
        "  return model\n",
        "\n",
        "all_hparam_possibilities = [\n",
        "    {\"concat_top_layer\": [0, 1], \"do_ave_pool\": [0, 1], \"conv_layer_params\": [((16, 3, 1),),  ((32, 3, 1), (16, 3, 2))], \"reg_amount\": [.1, .5, 1], \"drop_rate\": [0.0, .2, .4], \"fc_layer_sizes\": [(16,), (32,16),], 'reduction_alg': [None], 'n_components': [None]},\n",
        "]\n",
        "\n",
        "use_activations = False"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZpJ1xtu-7V9B",
        "colab_type": "code",
        "outputId": "38a48b22-c21e-4363-bdbc-d5fb8bbbf8f8",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "num_train = 50\n",
        "num_val = 1000\n",
        "epochs = 400\n",
        "print(\"use_activations:\", use_activations, \"num_train:\", num_train, \"epochs\", epochs)\n",
        "if num_train > 50:\n",
        "  print(\"More than 50 train data!!!!!!!!\")\n",
        "\n",
        "# each item in all_hparam_possibilities specifies valid hyper params to try. Put parameters that don't make sense together in separate lists.\n",
        "\n",
        "hparam_combinations = []\n",
        "for hparam_possibilities in all_hparam_possibilities:\n",
        "  hparam_keys, hparam_values = zip(*hparam_possibilities.items())\n",
        "  hparam_combinations.extend([dict(zip(hparam_keys, v)) for v in itertools.product(*hparam_values)])\n",
        "random.shuffle(hparam_combinations)\n",
        "print(\"len(hparam_combinations)\", len(hparam_combinations), \"hparam_combinations\", hparam_combinations)\n",
        "\n",
        "def modify_x_for_reduce(xs):\n",
        "  reshaped_x = np.reshape(xs, [xs.shape[0], -1])\n",
        "  # Make everything positive because some reductions don't work with negatives.\n",
        "  reshaped_x -= np.min(reshaped_x)\n",
        "  return reshaped_x\n",
        "\n",
        "def unsup_exstract(xs, reg_amount, drop_rate, layer_sizes, reduction_alg, n_components):\n",
        "  del reg_amount, drop_rate, layer_sizes\n",
        "\n",
        "  print(\"Using unsupervised feature extraction.\")\n",
        "\n",
        "  dim_reduct_model = ChannelReducer(reduction_alg=reduction_alg, n_components=n_components)\n",
        "  xs = dim_reduct_model.fit_transform(modify_x_for_reduce(xs))\n",
        "  return xs\n",
        "\n",
        "def convert_np(lst, i):\n",
        "  result = []\n",
        "  for x in lst:\n",
        "    result.append(x[i])\n",
        "  return np.asarray(result)\n",
        "\n",
        "def convert_xs(xs):\n",
        "  return {'img_input': convert_np(xs, 0), 'dense_input': convert_np(xs, 1)}\n",
        "\n",
        "def train_best_logs(xs, ys, num_val, do_summary, hparams, get_model):\n",
        "  \"\"\"Trains the model and retruns the logs of the best epoch. randomly splits the train and val data before training.\"\"\"\n",
        "  tf.keras.backend.clear_session()\n",
        "  model = get_model(**hparams)\n",
        "  xs, ys = shuffle(xs, ys)\n",
        "\n",
        "  xs_val = xs[num_train:num_train+num_val]\n",
        "  ys_val = ys[num_train:num_train+num_val]\n",
        "  early_stopping = tf.keras.callbacks.EarlyStopping(monitor='val_loss', patience=30, verbose=0)\n",
        "  best_stats = BestStats()\n",
        "  model.fit(convert_xs(xs[:num_train]), ys[:num_train], epochs=epochs, batch_size=256, validation_freq=1, callbacks=[best_stats, early_stopping], validation_data=(convert_xs(xs_val), ys_val), verbose=0)\n",
        "  if do_summary:\n",
        "    model.summary()\n",
        "    print(\"best train accuracy:\", best_stats.bestTrain)\n",
        "    print(\"Number of epochs:\", best_stats.num_epochs)\n",
        "  return best_stats.bestLogs\n",
        "\n",
        "def multiple_train_ave(hparams):\n",
        "  \"\"\"Trains the model multiple times with the same parameters and returns the average metrics\"\"\"\n",
        "  start = time.time()\n",
        "  all_val_auc = []\n",
        "  all_val_accuracy = []\n",
        "\n",
        "  if hparams['reduction_alg'] != None:\n",
        "    xs_for_train = unsup_exstract(xs, **hparams)\n",
        "  else:\n",
        "    xs_for_train = xs\n",
        "\n",
        "  do_summary = True\n",
        "  for i in range(5):\n",
        "    logs = train_best_logs(xs_for_train, ys, num_val, do_summary, hparams, get_model)\n",
        "    all_val_auc.append(get_val_auc(logs))\n",
        "    all_val_accuracy.append(logs.get('val_accuracy'))\n",
        "    do_summary = False \n",
        "\n",
        "  mean_val_auc = np.mean(all_val_auc)\n",
        "  mean_val_accuracy = np.mean(all_val_accuracy)\n",
        "  metric = (mean_val_auc + mean_val_accuracy) / 2.0\n",
        "  print_data = (\"mean_val_auc\", mean_val_auc, \"mean_val_accuracy\", mean_val_accuracy, \"metric\", metric, \"val_auc_std\", np.std(all_val_auc), \"val_accuracy_std\", np.std(all_val_accuracy))\n",
        "\n",
        "  end = time.time()\n",
        "  print(\"Seconds per hyperparam config\", end - start)\n",
        "  # GPU: ('Seconds per hyperparam config', 16.970870971679688)\n",
        "\n",
        "  return metric, print_data\n",
        "\n",
        "best_metric = -float('inf')\n",
        "\n",
        "run_num = 0\n",
        "for hparams in hparam_combinations:\n",
        "  print(\"hparams\", hparams)\n",
        "\n",
        "  metric, print_data = multiple_train_ave(hparams)\n",
        "\n",
        "  print(print_data)\n",
        "  if metric > best_metric:\n",
        "    best_metric = metric\n",
        "    best_print_data = print_data\n",
        "    best_hparams = hparams\n",
        "\n",
        "  run_num += 1\n",
        "  print(\"fract done\", run_num/float(len(hparam_combinations)))\n",
        "  print\n",
        "  print(\"==============================================================================================\")\n",
        "  print\n",
        "  sys.stdout.flush()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "use_activations: False num_train: 50 epochs 400\n",
            "len(hparam_combinations) 144 hparam_combinations [{'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}, {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}]\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            10817       sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 35,633\n",
            "Trainable params: 35,633\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 62\n",
            "Seconds per hyperparam config 26.2813618183136\n",
            "('mean_val_auc', 0.9166463732719421, 'mean_val_accuracy', 0.8597999930381774, 'metric', 0.8882231831550598, 'val_auc_std', 0.005964984819365684, 'val_accuracy_std', 0.009987997397410723)\n",
            "fract done 0.006944444444444444\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            3137        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 4,017\n",
            "Trainable params: 4,017\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.9399999976158142\n",
            "Number of epochs: 295\n",
            "Seconds per hyperparam config 51.19987511634827\n",
            "('mean_val_auc', 0.8525079250335693, 'mean_val_accuracy', 0.7597999930381775, 'metric', 0.8061539590358734, 'val_auc_std', 0.03196369577565304, 'val_accuracy_std', 0.09358717348274459)\n",
            "fract done 0.013888888888888888\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            289         sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 25,105\n",
            "Trainable params: 25,105\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 76\n",
            "Seconds per hyperparam config 25.94820523262024\n",
            "('mean_val_auc', 0.9075315952301025, 'mean_val_accuracy', 0.8533999919891357, 'metric', 0.8804657936096192, 'val_auc_std', 0.010087110328784172, 'val_accuracy_std', 0.008163321888619155)\n",
            "fract done 0.020833333333333332\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            3137        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 4,017\n",
            "Trainable params: 4,017\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.9800000190734863\n",
            "Number of epochs: 76\n",
            "Seconds per hyperparam config 27.436870098114014\n",
            "('mean_val_auc', 0.8751324772834778, 'mean_val_accuracy', 0.7984000086784363, 'metric', 0.836766242980957, 'val_auc_std', 0.026389266764056965, 'val_accuracy_std', 0.025523327919712777)\n",
            "fract done 0.027777777777777776\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1089        sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 11,185\n",
            "Trainable params: 11,185\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.9599999785423279\n",
            "Number of epochs: 75\n",
            "Seconds per hyperparam config 23.62440586090088\n",
            "('mean_val_auc', 0.9115562677383423, 'mean_val_accuracy', 0.8559999942779541, 'metric', 0.8837781310081483, 'val_auc_std', 0.012999441750241715, 'val_accuracy_std', 0.012521976538511765)\n",
            "fract done 0.034722222222222224\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 1984)         0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            31777       concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 32,657\n",
            "Trainable params: 32,657\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 77\n",
            "Seconds per hyperparam config 22.855773210525513\n",
            "('mean_val_auc', 0.8589723229408264, 'mean_val_accuracy', 0.8055999994277954, 'metric', 0.8322861611843109, 'val_auc_std', 0.03243749164037951, 'val_accuracy_std', 0.024548714501166403)\n",
            "fract done 0.041666666666666664\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 384)          0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            12865       concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 19,249\n",
            "Trainable params: 19,249\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 53\n",
            "Seconds per hyperparam config 19.13244938850403\n",
            "('mean_val_auc', 0.8308818697929382, 'mean_val_accuracy', 0.7659999966621399, 'metric', 0.7984409332275391, 'val_auc_std', 0.013145894915782212, 'val_accuracy_std', 0.030126384867898245)\n",
            "fract done 0.04861111111111111\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            30753       sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 40,849\n",
            "Trainable params: 40,849\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.9800000190734863\n",
            "Number of epochs: 89\n",
            "Seconds per hyperparam config 27.72665309906006\n",
            "('mean_val_auc', 0.8946720719337463, 'mean_val_accuracy', 0.8447999954223633, 'metric', 0.8697360336780549, 'val_auc_std', 0.014377787922045982, 'val_accuracy_std', 0.021609265044646563)\n",
            "fract done 0.05555555555555555\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            3137        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 9,521\n",
            "Trainable params: 9,521\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 53\n",
            "Seconds per hyperparam config 18.59333562850952\n",
            "('mean_val_auc', 0.8613149166107178, 'mean_val_accuracy', 0.7748000025749207, 'metric', 0.8180574595928192, 'val_auc_std', 0.028669996526286993, 'val_accuracy_std', 0.041019027021765804)\n",
            "fract done 0.0625\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1089        sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 11,185\n",
            "Trainable params: 11,185\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 99\n",
            "Seconds per hyperparam config 32.0841269493103\n",
            "('mean_val_auc', 0.9077357411384582, 'mean_val_accuracy', 0.8508000016212464, 'metric', 0.8792678713798523, 'val_auc_std', 0.021501274582900482, 'val_accuracy_std', 0.026483192469487857)\n",
            "fract done 0.06944444444444445\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1089        sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 25,905\n",
            "Trainable params: 25,905\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 71\n",
            "Seconds per hyperparam config 26.649015188217163\n",
            "('mean_val_auc', 0.9102641820907593, 'mean_val_accuracy', 0.8545999884605407, 'metric', 0.88243208527565, 'val_auc_std', 0.00642056373680674, 'val_accuracy_std', 0.02149976887157737)\n",
            "fract done 0.0763888888888889\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            289         sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 25,105\n",
            "Trainable params: 25,105\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 48\n",
            "Seconds per hyperparam config 18.256500720977783\n",
            "('mean_val_auc', 0.9093441367149353, 'mean_val_accuracy', 0.8552000045776367, 'metric', 0.882272070646286, 'val_auc_std', 0.010673656032787287, 'val_accuracy_std', 0.009662305723865084)\n",
            "fract done 0.08333333333333333\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            10817       sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 35,633\n",
            "Trainable params: 35,633\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 104\n",
            "Seconds per hyperparam config 25.898211002349854\n",
            "('mean_val_auc', 0.9034319043159484, 'mean_val_accuracy', 0.8546000123023987, 'metric', 0.8790159583091736, 'val_auc_std', 0.022187462808688083, 'val_accuracy_std', 0.02972944710385351)\n",
            "fract done 0.09027777777777778\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1313        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 7,697\n",
            "Trainable params: 7,697\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 46\n",
            "Seconds per hyperparam config 15.973385572433472\n",
            "('mean_val_auc', 0.8445493102073669, 'mean_val_accuracy', 0.7740000009536743, 'metric', 0.8092746555805206, 'val_auc_std', 0.02452378779451097, 'val_accuracy_std', 0.03745398861332048)\n",
            "fract done 0.09722222222222222\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            3137        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 9,521\n",
            "Trainable params: 9,521\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.9599999785423279\n",
            "Number of epochs: 135\n",
            "Seconds per hyperparam config 30.69206166267395\n",
            "('mean_val_auc', 0.8580388069152832, 'mean_val_accuracy', 0.7941999912261963, 'metric', 0.8261193990707398, 'val_auc_std', 0.024147474236645955, 'val_accuracy_std', 0.026194641629679213)\n",
            "fract done 0.10416666666666667\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            62017       sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 72,113\n",
            "Trainable params: 72,113\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 121\n",
            "Seconds per hyperparam config 37.915595293045044\n",
            "('mean_val_auc', 0.8948425769805908, 'mean_val_accuracy', 0.825600004196167, 'metric', 0.8602212905883788, 'val_auc_std', 0.019755888850567248, 'val_accuracy_std', 0.04483570610778006)\n",
            "fract done 0.1111111111111111\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            3137        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 9,521\n",
            "Trainable params: 9,521\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 81\n",
            "Seconds per hyperparam config 24.409552812576294\n",
            "('mean_val_auc', 0.8112486720085144, 'mean_val_accuracy', 0.7144000053405761, 'metric', 0.7628243386745452, 'val_auc_std', 0.057702920895694745, 'val_accuracy_std', 0.06406122054267449)\n",
            "fract done 0.11805555555555555\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            5153        sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 29,969\n",
            "Trainable params: 29,969\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 63\n",
            "Seconds per hyperparam config 23.947921752929688\n",
            "('mean_val_auc', 0.8684648871421814, 'mean_val_accuracy', 0.8055999994277954, 'metric', 0.8370324432849884, 'val_auc_std', 0.021418723712642074, 'val_accuracy_std', 0.039927933357657715)\n",
            "fract done 0.125\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            289         sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 25,105\n",
            "Trainable params: 25,105\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 109\n",
            "Seconds per hyperparam config 24.798198699951172\n",
            "('mean_val_auc', 0.8972028374671936, 'mean_val_accuracy', 0.8449999809265136, 'metric', 0.8711014091968536, 'val_auc_std', 0.02387277555049163, 'val_accuracy_std', 0.0186868956299101)\n",
            "fract done 0.13194444444444445\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            289         sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 10,385\n",
            "Trainable params: 10,385\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 80\n",
            "Seconds per hyperparam config 21.77985715866089\n",
            "('mean_val_auc', 0.9078459501266479, 'mean_val_accuracy', 0.848199999332428, 'metric', 0.8780229747295379, 'val_auc_std', 0.0069522004742641206, 'val_accuracy_std', 0.011070676010310518)\n",
            "fract done 0.1388888888888889\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1313        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 2,193\n",
            "Trainable params: 2,193\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.9399999976158142\n",
            "Number of epochs: 93\n",
            "Seconds per hyperparam config 22.170682430267334\n",
            "('mean_val_auc', 0.9065211653709412, 'mean_val_accuracy', 0.8424000024795533, 'metric', 0.8744605839252473, 'val_auc_std', 0.00833605257083165, 'val_accuracy_std', 0.007227730305781628)\n",
            "fract done 0.14583333333333334\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1089        sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 11,185\n",
            "Trainable params: 11,185\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 104\n",
            "Seconds per hyperparam config 33.45522856712341\n",
            "('mean_val_auc', 0.9054657816886902, 'mean_val_accuracy', 0.8313999891281127, 'metric', 0.8684328854084015, 'val_auc_std', 0.01128734125074317, 'val_accuracy_std', 0.015239416601180628)\n",
            "fract done 0.1527777777777778\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            10817       sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 35,633\n",
            "Trainable params: 35,633\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 77\n",
            "Seconds per hyperparam config 26.12491202354431\n",
            "('mean_val_auc', 0.9029217004776001, 'mean_val_accuracy', 0.8399999976158142, 'metric', 0.8714608490467072, 'val_auc_std', 0.025864309512143746, 'val_accuracy_std', 0.03855905282117897)\n",
            "fract done 0.1597222222222222\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1313        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 2,193\n",
            "Trainable params: 2,193\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.9800000190734863\n",
            "Number of epochs: 93\n",
            "Seconds per hyperparam config 28.959167003631592\n",
            "('mean_val_auc', 0.8689299702644349, 'mean_val_accuracy', 0.8038000106811524, 'metric', 0.8363649904727937, 'val_auc_std', 0.03202142575206936, 'val_accuracy_std', 0.03615741931129443)\n",
            "fract done 0.16666666666666666\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            10817       sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 35,633\n",
            "Trainable params: 35,633\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 138\n",
            "Seconds per hyperparam config 31.697550296783447\n",
            "('mean_val_auc', 0.906485378742218, 'mean_val_accuracy', 0.8470000147819519, 'metric', 0.8767426967620849, 'val_auc_std', 0.011659925389409732, 'val_accuracy_std', 0.010469004859672093)\n",
            "fract done 0.1736111111111111\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1089        sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 25,905\n",
            "Trainable params: 25,905\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 88\n",
            "Seconds per hyperparam config 27.763317823410034\n",
            "('mean_val_auc', 0.8749314069747924, 'mean_val_accuracy', 0.8191999912261962, 'metric', 0.8470656991004943, 'val_auc_std', 0.029450088543044064, 'val_accuracy_std', 0.022675099577264036)\n",
            "fract done 0.18055555555555555\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1313        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 7,697\n",
            "Trainable params: 7,697\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.9399999976158142\n",
            "Number of epochs: 56\n",
            "Seconds per hyperparam config 17.774078130722046\n",
            "('mean_val_auc', 0.819998836517334, 'mean_val_accuracy', 0.7151999950408936, 'metric', 0.7675994157791137, 'val_auc_std', 0.02969226718459965, 'val_accuracy_std', 0.056047828826915495)\n",
            "fract done 0.1875\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 384)          0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            6177        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 12,561\n",
            "Trainable params: 12,561\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 40\n",
            "Seconds per hyperparam config 14.324381351470947\n",
            "('mean_val_auc', 0.824114191532135, 'mean_val_accuracy', 0.7601999998092651, 'metric', 0.7921570956707, 'val_auc_std', 0.024532839599456208, 'val_accuracy_std', 0.01626530860615799)\n",
            "fract done 0.19444444444444445\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 1984)         0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            31777       concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 32,657\n",
            "Trainable params: 32,657\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 68\n",
            "Seconds per hyperparam config 19.694287300109863\n",
            "('mean_val_auc', 0.8224243283271789, 'mean_val_accuracy', 0.7575999975204468, 'metric', 0.7900121629238128, 'val_auc_std', 0.04547746077008363, 'val_accuracy_std', 0.044098064407381905)\n",
            "fract done 0.2013888888888889\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1089        sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 25,905\n",
            "Trainable params: 25,905\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 76\n",
            "Seconds per hyperparam config 27.040656328201294\n",
            "('mean_val_auc', 0.9207961320877075, 'mean_val_accuracy', 0.8561999917030334, 'metric', 0.8884980618953704, 'val_auc_std', 0.009590487916213405, 'val_accuracy_std', 0.01606734811491559)\n",
            "fract done 0.20833333333333334\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            5153        sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 29,969\n",
            "Trainable params: 29,969\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 39\n",
            "Seconds per hyperparam config 17.140939235687256\n",
            "('mean_val_auc', 0.8757688164710998, 'mean_val_accuracy', 0.8371999979019165, 'metric', 0.8564844071865081, 'val_auc_std', 0.01458545680042133, 'val_accuracy_std', 0.020922702063358913)\n",
            "fract done 0.2152777777777778\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 384)          0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            6177        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 12,561\n",
            "Trainable params: 12,561\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 41\n",
            "Seconds per hyperparam config 18.028516054153442\n",
            "('mean_val_auc', 0.8273806095123291, 'mean_val_accuracy', 0.7196000099182129, 'metric', 0.773490309715271, 'val_auc_std', 0.023144389759109887, 'val_accuracy_std', 0.06814279200442026)\n",
            "fract done 0.2222222222222222\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            3137        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 9,521\n",
            "Trainable params: 9,521\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.9599999785423279\n",
            "Number of epochs: 88\n",
            "Seconds per hyperparam config 24.317132711410522\n",
            "('mean_val_auc', 0.8363283753395081, 'mean_val_accuracy', 0.771399986743927, 'metric', 0.8038641810417175, 'val_auc_std', 0.022928050956500284, 'val_accuracy_std', 0.00717217282676503)\n",
            "fract done 0.22916666666666666\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 384)          0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            6177        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 12,561\n",
            "Trainable params: 12,561\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 40\n",
            "Seconds per hyperparam config 13.837090015411377\n",
            "('mean_val_auc', 0.82221599817276, 'mean_val_accuracy', 0.7516000032424927, 'metric', 0.7869080007076263, 'val_auc_std', 0.023910363112558237, 'val_accuracy_std', 0.015856849373403963)\n",
            "fract done 0.2361111111111111\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 1984)         0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            31777       concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 32,657\n",
            "Trainable params: 32,657\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 65\n",
            "Seconds per hyperparam config 23.786630868911743\n",
            "('mean_val_auc', 0.8627443909645081, 'mean_val_accuracy', 0.8058000087738038, 'metric', 0.8342721998691559, 'val_auc_std', 0.033431004654937466, 'val_accuracy_std', 0.04131537812103535)\n",
            "fract done 0.24305555555555555\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 1984)         0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            31777       concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 32,657\n",
            "Trainable params: 32,657\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 57\n",
            "Seconds per hyperparam config 18.317798614501953\n",
            "('mean_val_auc', 0.8381870985031128, 'mean_val_accuracy', 0.776800000667572, 'metric', 0.8074935495853424, 'val_auc_std', 0.035290177525841154, 'val_accuracy_std', 0.03916835789317768)\n",
            "fract done 0.25\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            289         sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 10,385\n",
            "Trainable params: 10,385\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 65\n",
            "Seconds per hyperparam config 21.174312591552734\n",
            "('mean_val_auc', 0.916353702545166, 'mean_val_accuracy', 0.8508000135421753, 'metric', 0.8835768580436707, 'val_auc_std', 0.006512419994577029, 'val_accuracy_std', 0.00738646294921255)\n",
            "fract done 0.2569444444444444\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            3137        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 9,521\n",
            "Trainable params: 9,521\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 60\n",
            "Seconds per hyperparam config 19.042536735534668\n",
            "('mean_val_auc', 0.844002902507782, 'mean_val_accuracy', 0.7608000040054321, 'metric', 0.8024014532566071, 'val_auc_std', 0.027549115406643315, 'val_accuracy_std', 0.035436707223575226)\n",
            "fract done 0.2638888888888889\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 1984)         0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            31777       concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 32,657\n",
            "Trainable params: 32,657\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 70\n",
            "Seconds per hyperparam config 19.96791648864746\n",
            "('mean_val_auc', 0.8315815448760986, 'mean_val_accuracy', 0.7662000060081482, 'metric', 0.7988907754421234, 'val_auc_std', 0.0475508077805861, 'val_accuracy_std', 0.04968460126797252)\n",
            "fract done 0.2708333333333333\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 384)          0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            6177        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 12,561\n",
            "Trainable params: 12,561\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 41\n",
            "Seconds per hyperparam config 14.462263822555542\n",
            "('mean_val_auc', 0.8482978820800782, 'mean_val_accuracy', 0.7784000039100647, 'metric', 0.8133489429950714, 'val_auc_std', 0.026470003194663347, 'val_accuracy_std', 0.030650283148177637)\n",
            "fract done 0.2777777777777778\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 384)          0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            6177        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 12,561\n",
            "Trainable params: 12,561\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 40\n",
            "Seconds per hyperparam config 13.994778394699097\n",
            "('mean_val_auc', 0.7977233529090881, 'mean_val_accuracy', 0.7336000084877015, 'metric', 0.7656616806983948, 'val_auc_std', 0.04001636841291769, 'val_accuracy_std', 0.05009830754594691)\n",
            "fract done 0.2847222222222222\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            10817       sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 35,633\n",
            "Trainable params: 35,633\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 87\n",
            "Seconds per hyperparam config 29.942402362823486\n",
            "('mean_val_auc', 0.8826260447502137, 'mean_val_accuracy', 0.8320000052452088, 'metric', 0.8573130249977112, 'val_auc_std', 0.044760365488585606, 'val_accuracy_std', 0.05015974997790058)\n",
            "fract done 0.2916666666666667\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            3137        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 9,521\n",
            "Trainable params: 9,521\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 86\n",
            "Seconds per hyperparam config 25.218832969665527\n",
            "('mean_val_auc', 0.8491260766983032, 'mean_val_accuracy', 0.7858000040054322, 'metric', 0.8174630403518677, 'val_auc_std', 0.028839739306694256, 'val_accuracy_std', 0.02793849442250833)\n",
            "fract done 0.2986111111111111\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            289         sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 25,105\n",
            "Trainable params: 25,105\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 84\n",
            "Seconds per hyperparam config 25.58363699913025\n",
            "('mean_val_auc', 0.8997101783752441, 'mean_val_accuracy', 0.8360000133514405, 'metric', 0.8678550958633423, 'val_auc_std', 0.013849429939623988, 'val_accuracy_std', 0.0186868956299101)\n",
            "fract done 0.3055555555555556\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            62017       sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 72,113\n",
            "Trainable params: 72,113\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.9599999785423279\n",
            "Number of epochs: 126\n",
            "Seconds per hyperparam config 34.10366439819336\n",
            "('mean_val_auc', 0.9026848673820496, 'mean_val_accuracy', 0.8424000024795533, 'metric', 0.8725424349308014, 'val_auc_std', 0.00911735860218437, 'val_accuracy_std', 0.00747261937741173)\n",
            "fract done 0.3125\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            289         sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 25,105\n",
            "Trainable params: 25,105\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 82\n",
            "Seconds per hyperparam config 25.122252225875854\n",
            "('mean_val_auc', 0.9033134818077088, 'mean_val_accuracy', 0.8383999943733216, 'metric', 0.8708567380905152, 'val_auc_std', 0.025423310628709707, 'val_accuracy_std', 0.026058401740799177)\n",
            "fract done 0.3194444444444444\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 1984)         0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            64065       concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 64,945\n",
            "Trainable params: 64,945\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 109\n",
            "Seconds per hyperparam config 27.929608583450317\n",
            "('mean_val_auc', 0.8451610922813415, 'mean_val_accuracy', 0.7863999962806701, 'metric', 0.8157805442810058, 'val_auc_std', 0.03805815864638014, 'val_accuracy_std', 0.04711518054481529)\n",
            "fract done 0.3263888888888889\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            30753       sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 40,849\n",
            "Trainable params: 40,849\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.9800000190734863\n",
            "Number of epochs: 76\n",
            "Seconds per hyperparam config 25.501798152923584\n",
            "('mean_val_auc', 0.8649785876274109, 'mean_val_accuracy', 0.8050000071525574, 'metric', 0.8349892973899842, 'val_auc_std', 0.03938713572536024, 'val_accuracy_std', 0.04085583341636599)\n",
            "fract done 0.3333333333333333\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            10817       sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 35,633\n",
            "Trainable params: 35,633\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.8799999952316284\n",
            "Number of epochs: 129\n",
            "Seconds per hyperparam config 28.601557970046997\n",
            "('mean_val_auc', 0.8685070753097535, 'mean_val_accuracy', 0.8126000165939331, 'metric', 0.8405535459518433, 'val_auc_std', 0.02119289130660933, 'val_accuracy_std', 0.025421237663438512)\n",
            "fract done 0.3402777777777778\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            30753       sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 40,849\n",
            "Trainable params: 40,849\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 72\n",
            "Seconds per hyperparam config 23.96021008491516\n",
            "('mean_val_auc', 0.8844434142112731, 'mean_val_accuracy', 0.8162000060081482, 'metric', 0.8503217101097107, 'val_auc_std', 0.033951014278870435, 'val_accuracy_std', 0.0331927683934229)\n",
            "fract done 0.3472222222222222\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1089        sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 11,185\n",
            "Trainable params: 11,185\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.9800000190734863\n",
            "Number of epochs: 76\n",
            "Seconds per hyperparam config 29.663002490997314\n",
            "('mean_val_auc', 0.8983074307441712, 'mean_val_accuracy', 0.8202000021934509, 'metric', 0.859253716468811, 'val_auc_std', 0.00919647013347605, 'val_accuracy_std', 0.019239533956487424)\n",
            "fract done 0.3541666666666667\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            3137        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 9,521\n",
            "Trainable params: 9,521\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.9800000190734863\n",
            "Number of epochs: 50\n",
            "Seconds per hyperparam config 18.761625051498413\n",
            "('mean_val_auc', 0.8583778500556946, 'mean_val_accuracy', 0.7651999950408935, 'metric', 0.811788922548294, 'val_auc_std', 0.041081153360688356, 'val_accuracy_std', 0.05701016107005085)\n",
            "fract done 0.3611111111111111\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1089        sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 11,185\n",
            "Trainable params: 11,185\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 73\n",
            "Seconds per hyperparam config 22.344246864318848\n",
            "('mean_val_auc', 0.9206916213035583, 'mean_val_accuracy', 0.850600004196167, 'metric', 0.8856458127498626, 'val_auc_std', 0.01818748913122844, 'val_accuracy_std', 0.026507362303170343)\n",
            "fract done 0.3680555555555556\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            5153        sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 29,969\n",
            "Trainable params: 29,969\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 82\n",
            "Seconds per hyperparam config 26.21631145477295\n",
            "('mean_val_auc', 0.8947139024734497, 'mean_val_accuracy', 0.848199987411499, 'metric', 0.8714569449424744, 'val_auc_std', 0.010638664553431284, 'val_accuracy_std', 0.017656728315783708)\n",
            "fract done 0.375\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            5153        sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 29,969\n",
            "Trainable params: 29,969\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 84\n",
            "Seconds per hyperparam config 23.50100827217102\n",
            "('mean_val_auc', 0.9085950493812561, 'mean_val_accuracy', 0.8428000092506409, 'metric', 0.8756975293159486, 'val_auc_std', 0.013659432330558267, 'val_accuracy_std', 0.014372196303576348)\n",
            "fract done 0.3819444444444444\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 384)          0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            12865       concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 19,249\n",
            "Trainable params: 19,249\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 79\n",
            "Seconds per hyperparam config 27.618260145187378\n",
            "('mean_val_auc', 0.8227193713188171, 'mean_val_accuracy', 0.7379999995231629, 'metric', 0.78035968542099, 'val_auc_std', 0.01649510012184946, 'val_accuracy_std', 0.045821393016621075)\n",
            "fract done 0.3888888888888889\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 384)          0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            12865       concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 19,249\n",
            "Trainable params: 19,249\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 66\n",
            "Seconds per hyperparam config 22.034642696380615\n",
            "('mean_val_auc', 0.8177564263343811, 'mean_val_accuracy', 0.7263999938964844, 'metric', 0.7720782101154327, 'val_auc_std', 0.035190720289460936, 'val_accuracy_std', 0.044301693440485804)\n",
            "fract done 0.3958333333333333\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            289         sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 10,385\n",
            "Trainable params: 10,385\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 77\n",
            "Seconds per hyperparam config 24.34257435798645\n",
            "('mean_val_auc', 0.8975362181663513, 'mean_val_accuracy', 0.8340000033378601, 'metric', 0.8657681107521057, 'val_auc_std', 0.02290109968303664, 'val_accuracy_std', 0.02511571348526105)\n",
            "fract done 0.4027777777777778\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1313        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 7,697\n",
            "Trainable params: 7,697\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 54\n",
            "Seconds per hyperparam config 17.12056541442871\n",
            "('mean_val_auc', 0.8252241730690002, 'mean_val_accuracy', 0.7377999901771546, 'metric', 0.7815120816230774, 'val_auc_std', 0.03814770468371583, 'val_accuracy_std', 0.06577354129984979)\n",
            "fract done 0.4097222222222222\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            3137        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 4,017\n",
            "Trainable params: 4,017\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.9800000190734863\n",
            "Number of epochs: 204\n",
            "Seconds per hyperparam config 46.90776562690735\n",
            "('mean_val_auc', 0.8668850064277649, 'mean_val_accuracy', 0.8047999978065491, 'metric', 0.8358425021171569, 'val_auc_std', 0.03223656064897096, 'val_accuracy_std', 0.03469524247978946)\n",
            "fract done 0.4166666666666667\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            30753       sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 40,849\n",
            "Trainable params: 40,849\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 92\n",
            "Seconds per hyperparam config 26.90712594985962\n",
            "('mean_val_auc', 0.8592967629432678, 'mean_val_accuracy', 0.7810000061988831, 'metric', 0.8201483845710755, 'val_auc_std', 0.03152954927446335, 'val_accuracy_std', 0.03542315149485199)\n",
            "fract done 0.4236111111111111\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1313        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 7,697\n",
            "Trainable params: 7,697\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.9800000190734863\n",
            "Number of epochs: 56\n",
            "Seconds per hyperparam config 18.385835647583008\n",
            "('mean_val_auc', 0.8553674936294555, 'mean_val_accuracy', 0.7773999810218811, 'metric', 0.8163837373256684, 'val_auc_std', 0.02308323284269413, 'val_accuracy_std', 0.043361730747432985)\n",
            "fract done 0.4305555555555556\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 1984)         0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            64065       concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 64,945\n",
            "Trainable params: 64,945\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 82\n",
            "Seconds per hyperparam config 24.17221474647522\n",
            "('mean_val_auc', 0.8586001753807068, 'mean_val_accuracy', 0.8002000093460083, 'metric', 0.8294000923633575, 'val_auc_std', 0.020782191850987675, 'val_accuracy_std', 0.02230156883743404)\n",
            "fract done 0.4375\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1313        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 2,193\n",
            "Trainable params: 2,193\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 110\n",
            "Seconds per hyperparam config 30.57163643836975\n",
            "('mean_val_auc', 0.886538028717041, 'mean_val_accuracy', 0.8179999828338623, 'metric', 0.8522690057754516, 'val_auc_std', 0.016685704086728873, 'val_accuracy_std', 0.01856878563908093)\n",
            "fract done 0.4444444444444444\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 1984)         0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            64065       concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 64,945\n",
            "Trainable params: 64,945\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 103\n",
            "Seconds per hyperparam config 26.81366515159607\n",
            "('mean_val_auc', 0.8480786204338073, 'mean_val_accuracy', 0.7862000107765198, 'metric', 0.8171393156051636, 'val_auc_std', 0.021708300641433458, 'val_accuracy_std', 0.02185771466813769)\n",
            "fract done 0.4513888888888889\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            5153        sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 29,969\n",
            "Trainable params: 29,969\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 41\n",
            "Seconds per hyperparam config 19.7011981010437\n",
            "('mean_val_auc', 0.8783175230026246, 'mean_val_accuracy', 0.8131999969482422, 'metric', 0.8457587599754334, 'val_auc_std', 0.02322380700158912, 'val_accuracy_std', 0.023911491492425045)\n",
            "fract done 0.4583333333333333\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            30753       sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 40,849\n",
            "Trainable params: 40,849\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.9800000190734863\n",
            "Number of epochs: 82\n",
            "Seconds per hyperparam config 25.67294955253601\n",
            "('mean_val_auc', 0.8905142664909362, 'mean_val_accuracy', 0.8326000094413757, 'metric', 0.861557137966156, 'val_auc_std', 0.02234965473417519, 'val_accuracy_std', 0.02360169000393857)\n",
            "fract done 0.4652777777777778\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1313        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 7,697\n",
            "Trainable params: 7,697\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 48\n",
            "Seconds per hyperparam config 16.68311357498169\n",
            "('mean_val_auc', 0.8253544330596924, 'mean_val_accuracy', 0.7466000080108642, 'metric', 0.7859772205352783, 'val_auc_std', 0.05018388000881082, 'val_accuracy_std', 0.039174466393012516)\n",
            "fract done 0.4722222222222222\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1313        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 7,697\n",
            "Trainable params: 7,697\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 55\n",
            "Seconds per hyperparam config 18.013044595718384\n",
            "('mean_val_auc', 0.8374431610107422, 'mean_val_accuracy', 0.7489999890327453, 'metric', 0.7932215750217437, 'val_auc_std', 0.024259001407512696, 'val_accuracy_std', 0.04302092850398433)\n",
            "fract done 0.4791666666666667\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1313        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 2,193\n",
            "Trainable params: 2,193\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.9599999785423279\n",
            "Number of epochs: 113\n",
            "Seconds per hyperparam config 29.720426321029663\n",
            "('mean_val_auc', 0.8550212025642395, 'mean_val_accuracy', 0.771399998664856, 'metric', 0.8132106006145478, 'val_auc_std', 0.04404489482447431, 'val_accuracy_std', 0.06884359382102669)\n",
            "fract done 0.4861111111111111\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1089        sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 11,185\n",
            "Trainable params: 11,185\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.9800000190734863\n",
            "Number of epochs: 110\n",
            "Seconds per hyperparam config 39.26630425453186\n",
            "('mean_val_auc', 0.8919551610946655, 'mean_val_accuracy', 0.8236000061035156, 'metric', 0.8577775835990906, 'val_auc_std', 0.024985719758897667, 'val_accuracy_std', 0.043701711405102484)\n",
            "fract done 0.4930555555555556\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1313        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 2,193\n",
            "Trainable params: 2,193\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 266\n",
            "Seconds per hyperparam config 28.50161600112915\n",
            "('mean_val_auc', 0.8827928900718689, 'mean_val_accuracy', 0.7912000060081482, 'metric', 0.8369964480400085, 'val_auc_std', 0.03121512018542097, 'val_accuracy_std', 0.06473761836378428)\n",
            "fract done 0.5\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            289         sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 10,385\n",
            "Trainable params: 10,385\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 78\n",
            "Seconds per hyperparam config 25.031707763671875\n",
            "('mean_val_auc', 0.9128099322319031, 'mean_val_accuracy', 0.8559999942779541, 'metric', 0.8844049632549287, 'val_auc_std', 0.0071958711191259265, 'val_accuracy_std', 0.008671783435800247)\n",
            "fract done 0.5069444444444444\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            62017       sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 72,113\n",
            "Trainable params: 72,113\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 107\n",
            "Seconds per hyperparam config 30.23396062850952\n",
            "('mean_val_auc', 0.8856058597564698, 'mean_val_accuracy', 0.8227999925613403, 'metric', 0.854202926158905, 'val_auc_std', 0.031924371657498715, 'val_accuracy_std', 0.04741897905424655)\n",
            "fract done 0.5138888888888888\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1313        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 2,193\n",
            "Trainable params: 2,193\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 84\n",
            "Seconds per hyperparam config 29.892322778701782\n",
            "('mean_val_auc', 0.8873034834861755, 'mean_val_accuracy', 0.8104000091552734, 'metric', 0.8488517463207245, 'val_auc_std', 0.010110945000007216, 'val_accuracy_std', 0.007144214217916867)\n",
            "fract done 0.5208333333333334\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 1984)         0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            31777       concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 32,657\n",
            "Trainable params: 32,657\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 83\n",
            "Seconds per hyperparam config 23.123594045639038\n",
            "('mean_val_auc', 0.8741771817207337, 'mean_val_accuracy', 0.8202000021934509, 'metric', 0.8471885919570923, 'val_auc_std', 0.022373096594736352, 'val_accuracy_std', 0.02360847600976517)\n",
            "fract done 0.5277777777777778\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            289         sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 10,385\n",
            "Trainable params: 10,385\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 74\n",
            "Seconds per hyperparam config 21.7955060005188\n",
            "('mean_val_auc', 0.9099945664405823, 'mean_val_accuracy', 0.846400010585785, 'metric', 0.8781972885131837, 'val_auc_std', 0.011014844953914565, 'val_accuracy_std', 0.02147185483823247)\n",
            "fract done 0.5347222222222222\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            62017       sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 72,113\n",
            "Trainable params: 72,113\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 135\n",
            "Seconds per hyperparam config 33.46616458892822\n",
            "('mean_val_auc', 0.8895633697509766, 'mean_val_accuracy', 0.8450000047683716, 'metric', 0.867281687259674, 'val_auc_std', 0.03038174485792625, 'val_accuracy_std', 0.03126021497170947)\n",
            "fract done 0.5416666666666666\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            30753       sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 40,849\n",
            "Trainable params: 40,849\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 94\n",
            "Seconds per hyperparam config 24.78055477142334\n",
            "('mean_val_auc', 0.8587526559829712, 'mean_val_accuracy', 0.7861999869346619, 'metric', 0.8224763214588166, 'val_auc_std', 0.0628173600361051, 'val_accuracy_std', 0.05508321638528817)\n",
            "fract done 0.5486111111111112\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            3137        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 9,521\n",
            "Trainable params: 9,521\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 90\n",
            "Seconds per hyperparam config 27.162712812423706\n",
            "('mean_val_auc', 0.7990536451339721, 'mean_val_accuracy', 0.7178000092506409, 'metric', 0.7584268271923065, 'val_auc_std', 0.04307077073330452, 'val_accuracy_std', 0.04936152365002281)\n",
            "fract done 0.5555555555555556\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            289         sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 10,385\n",
            "Trainable params: 10,385\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 84\n",
            "Seconds per hyperparam config 23.89621639251709\n",
            "('mean_val_auc', 0.8935268521308899, 'mean_val_accuracy', 0.825600004196167, 'metric', 0.8595634281635285, 'val_auc_std', 0.02723037645570683, 'val_accuracy_std', 0.04276727984091531)\n",
            "fract done 0.5625\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 384)          0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            6177        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 12,561\n",
            "Trainable params: 12,561\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 41\n",
            "Seconds per hyperparam config 16.043384552001953\n",
            "('mean_val_auc', 0.8235596776008606, 'mean_val_accuracy', 0.7652000069618226, 'metric', 0.7943798422813415, 'val_auc_std', 0.02982974744252225, 'val_accuracy_std', 0.023760470309762742)\n",
            "fract done 0.5694444444444444\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            30753       sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 40,849\n",
            "Trainable params: 40,849\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 84\n",
            "Seconds per hyperparam config 28.610477209091187\n",
            "('mean_val_auc', 0.8762506246566772, 'mean_val_accuracy', 0.8186000108718872, 'metric', 0.8474253177642822, 'val_auc_std', 0.02328522355678813, 'val_accuracy_std', 0.024204133007915338)\n",
            "fract done 0.5763888888888888\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1089        sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 25,905\n",
            "Trainable params: 25,905\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 66\n",
            "Seconds per hyperparam config 25.234687328338623\n",
            "('mean_val_auc', 0.9034214615821838, 'mean_val_accuracy', 0.84339998960495, 'metric', 0.8734107255935669, 'val_auc_std', 0.0173355523857128, 'val_accuracy_std', 0.020655252530349413)\n",
            "fract done 0.5833333333333334\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 1984)         0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            64065       concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 64,945\n",
            "Trainable params: 64,945\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 102\n",
            "Seconds per hyperparam config 24.14628553390503\n",
            "('mean_val_auc', 0.8372727513313294, 'mean_val_accuracy', 0.7807999968528747, 'metric', 0.809036374092102, 'val_auc_std', 0.06305428522049328, 'val_accuracy_std', 0.06182037845033109)\n",
            "fract done 0.5902777777777778\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 1984)         0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            31777       concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 32,657\n",
            "Trainable params: 32,657\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 77\n",
            "Seconds per hyperparam config 20.633612871170044\n",
            "('mean_val_auc', 0.8608517527580262, 'mean_val_accuracy', 0.7937999963760376, 'metric', 0.8273258745670319, 'val_auc_std', 0.022735452639726508, 'val_accuracy_std', 0.033017583004063905)\n",
            "fract done 0.5972222222222222\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            289         sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 10,385\n",
            "Trainable params: 10,385\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 83\n",
            "Seconds per hyperparam config 23.9360294342041\n",
            "('mean_val_auc', 0.9045521497726441, 'mean_val_accuracy', 0.8378000020980835, 'metric', 0.8711760759353637, 'val_auc_std', 0.012170913661017628, 'val_accuracy_std', 0.0209799919598998)\n",
            "fract done 0.6041666666666666\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 384)          0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            12865       concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 19,249\n",
            "Trainable params: 19,249\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 80\n",
            "Seconds per hyperparam config 22.592795610427856\n",
            "('mean_val_auc', 0.8143821001052857, 'mean_val_accuracy', 0.726800000667572, 'metric', 0.7705910503864288, 'val_auc_std', 0.03511806272839999, 'val_accuracy_std', 0.044638105668551406)\n",
            "fract done 0.6111111111111112\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            5153        sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 29,969\n",
            "Trainable params: 29,969\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.9800000190734863\n",
            "Number of epochs: 64\n",
            "Seconds per hyperparam config 16.400784015655518\n",
            "('mean_val_auc', 0.9065890073776245, 'mean_val_accuracy', 0.8365999937057496, 'metric', 0.871594500541687, 'val_auc_std', 0.01673259399005325, 'val_accuracy_std', 0.019011570307714432)\n",
            "fract done 0.6180555555555556\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1089        sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 25,905\n",
            "Trainable params: 25,905\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 101\n",
            "Seconds per hyperparam config 30.701151371002197\n",
            "('mean_val_auc', 0.9061676025390625, 'mean_val_accuracy', 0.8437999963760376, 'metric', 0.87498379945755, 'val_auc_std', 0.012920712534584475, 'val_accuracy_std', 0.01539348908056043)\n",
            "fract done 0.625\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            62017       sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 72,113\n",
            "Trainable params: 72,113\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 107\n",
            "Seconds per hyperparam config 29.977526426315308\n",
            "('mean_val_auc', 0.9108912467956543, 'mean_val_accuracy', 0.854799997806549, 'metric', 0.8828456223011016, 'val_auc_std', 0.012772887664236298, 'val_accuracy_std', 0.010438399729080299)\n",
            "fract done 0.6319444444444444\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 1984)         0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            31777       concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 32,657\n",
            "Trainable params: 32,657\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 96\n",
            "Seconds per hyperparam config 27.04892873764038\n",
            "('mean_val_auc', 0.8569633007049561, 'mean_val_accuracy', 0.7930000066757202, 'metric', 0.8249816536903382, 'val_auc_std', 0.016056528948414354, 'val_accuracy_std', 0.019798980625005406)\n",
            "fract done 0.6388888888888888\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1313        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 2,193\n",
            "Trainable params: 2,193\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.9800000190734863\n",
            "Number of epochs: 114\n",
            "Seconds per hyperparam config 30.72835683822632\n",
            "('mean_val_auc', 0.8778395771980285, 'mean_val_accuracy', 0.8128000020980835, 'metric', 0.845319789648056, 'val_auc_std', 0.037418468268423266, 'val_accuracy_std', 0.04999360490566761)\n",
            "fract done 0.6458333333333334\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            289         sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 25,105\n",
            "Trainable params: 25,105\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 63\n",
            "Seconds per hyperparam config 20.350324630737305\n",
            "('mean_val_auc', 0.9116104364395141, 'mean_val_accuracy', 0.8549999952316284, 'metric', 0.8833052158355712, 'val_auc_std', 0.012976741541581042, 'val_accuracy_std', 0.010639545214484136)\n",
            "fract done 0.6527777777777778\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            10817       sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 35,633\n",
            "Trainable params: 35,633\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 82\n",
            "Seconds per hyperparam config 27.87047815322876\n",
            "('mean_val_auc', 0.906226634979248, 'mean_val_accuracy', 0.8453999996185303, 'metric', 0.8758133172988891, 'val_auc_std', 0.005269479570279494, 'val_accuracy_std', 0.0063119006607168515)\n",
            "fract done 0.6597222222222222\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            289         sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 10,385\n",
            "Trainable params: 10,385\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.9800000190734863\n",
            "Number of epochs: 81\n",
            "Seconds per hyperparam config 26.144832611083984\n",
            "('mean_val_auc', 0.9087708592414856, 'mean_val_accuracy', 0.8589999914169312, 'metric', 0.8838854253292083, 'val_auc_std', 0.025126195949015554, 'val_accuracy_std', 0.028982739014229177)\n",
            "fract done 0.6666666666666666\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 384)          0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            6177        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 12,561\n",
            "Trainable params: 12,561\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 79\n",
            "Seconds per hyperparam config 21.295114517211914\n",
            "('mean_val_auc', 0.816062617301941, 'mean_val_accuracy', 0.7437999963760376, 'metric', 0.7799313068389893, 'val_auc_std', 0.023685534275051046, 'val_accuracy_std', 0.037722148297062504)\n",
            "fract done 0.6736111111111112\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 384)          0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            6177        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 12,561\n",
            "Trainable params: 12,561\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 94\n",
            "Seconds per hyperparam config 20.154354572296143\n",
            "('mean_val_auc', 0.8079568982124329, 'mean_val_accuracy', 0.7435999989509583, 'metric', 0.7757784485816956, 'val_auc_std', 0.04100222635722972, 'val_accuracy_std', 0.032903493787319056)\n",
            "fract done 0.6805555555555556\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            10817       sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 35,633\n",
            "Trainable params: 35,633\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 93\n",
            "Seconds per hyperparam config 29.692359924316406\n",
            "('mean_val_auc', 0.9060781478881836, 'mean_val_accuracy', 0.8453999996185303, 'metric', 0.875739073753357, 'val_auc_std', 0.023907402541479453, 'val_accuracy_std', 0.02778920730171664)\n",
            "fract done 0.6875\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1313        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 2,193\n",
            "Trainable params: 2,193\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.9800000190734863\n",
            "Number of epochs: 101\n",
            "Seconds per hyperparam config 32.45576572418213\n",
            "('mean_val_auc', 0.8927137851715088, 'mean_val_accuracy', 0.8255999922752381, 'metric', 0.8591568887233734, 'val_auc_std', 0.02197633271693073, 'val_accuracy_std', 0.036603820498717424)\n",
            "fract done 0.6944444444444444\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 384)          0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            12865       concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 19,249\n",
            "Trainable params: 19,249\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 93\n",
            "Seconds per hyperparam config 27.120441436767578\n",
            "('mean_val_auc', 0.8076894760131836, 'mean_val_accuracy', 0.7460000038146972, 'metric', 0.7768447399139404, 'val_auc_std', 0.03782904447403953, 'val_accuracy_std', 0.026260233481072922)\n",
            "fract done 0.7013888888888888\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            5153        sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 29,969\n",
            "Trainable params: 29,969\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 71\n",
            "Seconds per hyperparam config 21.542237043380737\n",
            "('mean_val_auc', 0.8978726267814636, 'mean_val_accuracy', 0.8445999979972839, 'metric', 0.8712363123893738, 'val_auc_std', 0.026653464093684447, 'val_accuracy_std', 0.02305299012987664)\n",
            "fract done 0.7083333333333334\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            3137        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 4,017\n",
            "Trainable params: 4,017\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.9399999976158142\n",
            "Number of epochs: 124\n",
            "Seconds per hyperparam config 40.452561378479004\n",
            "('mean_val_auc', 0.8616438984870911, 'mean_val_accuracy', 0.7808000087738037, 'metric', 0.8212219536304475, 'val_auc_std', 0.03788051186319581, 'val_accuracy_std', 0.04466049032657429)\n",
            "fract done 0.7152777777777778\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1089        sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 11,185\n",
            "Trainable params: 11,185\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.9800000190734863\n",
            "Number of epochs: 125\n",
            "Seconds per hyperparam config 29.442238569259644\n",
            "('mean_val_auc', 0.9142109990119934, 'mean_val_accuracy', 0.8658000111579895, 'metric', 0.8900055050849915, 'val_auc_std', 0.011914377924630034, 'val_accuracy_std', 0.014743140828623479)\n",
            "fract done 0.7222222222222222\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            3137        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 4,017\n",
            "Trainable params: 4,017\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.9800000190734863\n",
            "Number of epochs: 139\n",
            "Seconds per hyperparam config 30.026965379714966\n",
            "('mean_val_auc', 0.9005542516708374, 'mean_val_accuracy', 0.8314000010490418, 'metric', 0.8659771263599396, 'val_auc_std', 0.024495804654865304, 'val_accuracy_std', 0.016304601084303424)\n",
            "fract done 0.7291666666666666\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            3137        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 4,017\n",
            "Trainable params: 4,017\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.9200000166893005\n",
            "Number of epochs: 102\n",
            "Seconds per hyperparam config 48.58560109138489\n",
            "('mean_val_auc', 0.8683247208595276, 'mean_val_accuracy', 0.770200002193451, 'metric', 0.8192623615264893, 'val_auc_std', 0.0270701358077647, 'val_accuracy_std', 0.0618850527081297)\n",
            "fract done 0.7361111111111112\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            3137        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 4,017\n",
            "Trainable params: 4,017\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.9399999976158142\n",
            "Number of epochs: 243\n",
            "Seconds per hyperparam config 55.891382455825806\n",
            "('mean_val_auc', 0.8706436634063721, 'mean_val_accuracy', 0.7728000164031983, 'metric', 0.8217218399047852, 'val_auc_std', 0.019505011275441134, 'val_accuracy_std', 0.08049198968518909)\n",
            "fract done 0.7430555555555556\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 1984)         0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            64065       concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 64,945\n",
            "Trainable params: 64,945\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 133\n",
            "Seconds per hyperparam config 32.37525177001953\n",
            "('mean_val_auc', 0.8281836867332458, 'mean_val_accuracy', 0.7621999979019165, 'metric', 0.7951918423175812, 'val_auc_std', 0.03107584556291333, 'val_accuracy_std', 0.032399999874614525)\n",
            "fract done 0.75\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            62017       sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 72,113\n",
            "Trainable params: 72,113\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.9800000190734863\n",
            "Number of epochs: 83\n",
            "Seconds per hyperparam config 29.499386072158813\n",
            "('mean_val_auc', 0.8812717914581298, 'mean_val_accuracy', 0.8301999926567077, 'metric', 0.8557358920574187, 'val_auc_std', 0.041980714555953556, 'val_accuracy_std', 0.04200189606428858)\n",
            "fract done 0.7569444444444444\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            3137        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 4,017\n",
            "Trainable params: 4,017\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 147\n",
            "Seconds per hyperparam config 31.271673440933228\n",
            "('mean_val_auc', 0.909014880657196, 'mean_val_accuracy', 0.8465999960899353, 'metric', 0.8778074383735657, 'val_auc_std', 0.02289967983896212, 'val_accuracy_std', 0.029179452191941988)\n",
            "fract done 0.7638888888888888\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            62017       sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 72,113\n",
            "Trainable params: 72,113\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 113\n",
            "Seconds per hyperparam config 30.695661067962646\n",
            "('mean_val_auc', 0.862993848323822, 'mean_val_accuracy', 0.8057999968528747, 'metric', 0.8343969225883483, 'val_auc_std', 0.037120368710069705, 'val_accuracy_std', 0.017336665242744857)\n",
            "fract done 0.7708333333333334\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1313        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 7,697\n",
            "Trainable params: 7,697\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 57\n",
            "Seconds per hyperparam config 17.819671392440796\n",
            "('mean_val_auc', 0.8308543205261231, 'mean_val_accuracy', 0.7582000017166137, 'metric', 0.7945271611213685, 'val_auc_std', 0.026899521426815638, 'val_accuracy_std', 0.03296300042719972)\n",
            "fract done 0.7777777777777778\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 384)          0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            12865       concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 19,249\n",
            "Trainable params: 19,249\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 76\n",
            "Seconds per hyperparam config 27.608936548233032\n",
            "('mean_val_auc', 0.828579556941986, 'mean_val_accuracy', 0.7656000018119812, 'metric', 0.7970897793769836, 'val_auc_std', 0.04310331975964626, 'val_accuracy_std', 0.04532815241805659)\n",
            "fract done 0.7847222222222222\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            289         sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 25,105\n",
            "Trainable params: 25,105\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 84\n",
            "Seconds per hyperparam config 23.975878715515137\n",
            "('mean_val_auc', 0.9123581290245056, 'mean_val_accuracy', 0.8559999942779541, 'metric', 0.88417906165123, 'val_auc_std', 0.022223009952782195, 'val_accuracy_std', 0.0202780598581067)\n",
            "fract done 0.7916666666666666\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 1984)         0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            64065       concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 64,945\n",
            "Trainable params: 64,945\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 70\n",
            "Seconds per hyperparam config 21.88998007774353\n",
            "('mean_val_auc', 0.8139037966728211, 'mean_val_accuracy', 0.7606000065803528, 'metric', 0.7872519016265869, 'val_auc_std', 0.08596578800659654, 'val_accuracy_std', 0.05547827237909979)\n",
            "fract done 0.7986111111111112\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            289         sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 25,105\n",
            "Trainable params: 25,105\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 68\n",
            "Seconds per hyperparam config 25.026474237442017\n",
            "('mean_val_auc', 0.9066962957382202, 'mean_val_accuracy', 0.8400000214576722, 'metric', 0.8733481585979461, 'val_auc_std', 0.009435389983537308, 'val_accuracy_std', 0.014099642253964297)\n",
            "fract done 0.8055555555555556\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1313        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 7,697\n",
            "Trainable params: 7,697\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 47\n",
            "Seconds per hyperparam config 15.422330856323242\n",
            "('mean_val_auc', 0.8341182112693787, 'mean_val_accuracy', 0.751199996471405, 'metric', 0.7926591038703918, 'val_auc_std', 0.027906575806576978, 'val_accuracy_std', 0.0480682764284237)\n",
            "fract done 0.8125\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 384)          0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            12865       concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 19,249\n",
            "Trainable params: 19,249\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 67\n",
            "Seconds per hyperparam config 21.09554696083069\n",
            "('mean_val_auc', 0.8414564490318298, 'mean_val_accuracy', 0.7696000099182129, 'metric', 0.8055282294750213, 'val_auc_std', 0.024389805062165256, 'val_accuracy_std', 0.028973078704674287)\n",
            "fract done 0.8194444444444444\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1089        sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 25,905\n",
            "Trainable params: 25,905\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 71\n",
            "Seconds per hyperparam config 27.936102867126465\n",
            "('mean_val_auc', 0.8948008060455322, 'mean_val_accuracy', 0.8450000047683716, 'metric', 0.8699004054069519, 'val_auc_std', 0.01981398864877806, 'val_accuracy_std', 0.022208106275863473)\n",
            "fract done 0.8263888888888888\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            5153        sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 29,969\n",
            "Trainable params: 29,969\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 39\n",
            "Seconds per hyperparam config 15.32908320426941\n",
            "('mean_val_auc', 0.8864269614219665, 'mean_val_accuracy', 0.8425999999046325, 'metric', 0.8645134806632995, 'val_auc_std', 0.012355179682959006, 'val_accuracy_std', 0.01693043357597378)\n",
            "fract done 0.8333333333333334\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1089        sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 11,185\n",
            "Trainable params: 11,185\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 102\n",
            "Seconds per hyperparam config 35.56260061264038\n",
            "('mean_val_auc', 0.893548595905304, 'mean_val_accuracy', 0.8296000003814697, 'metric', 0.8615742981433869, 'val_auc_std', 0.021493309570970297, 'val_accuracy_std', 0.036603838410809886)\n",
            "fract done 0.8402777777777778\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 1984)         0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            64065       concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 64,945\n",
            "Trainable params: 64,945\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 110\n",
            "Seconds per hyperparam config 29.696337938308716\n",
            "('mean_val_auc', 0.8547765135765075, 'mean_val_accuracy', 0.7934000015258789, 'metric', 0.8240882575511932, 'val_auc_std', 0.01670748996205691, 'val_accuracy_std', 0.027767623545397305)\n",
            "fract done 0.8472222222222222\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            289         sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 10,385\n",
            "Trainable params: 10,385\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 75\n",
            "Seconds per hyperparam config 23.20430064201355\n",
            "('mean_val_auc', 0.9092634201049805, 'mean_val_accuracy', 0.8549999952316284, 'metric', 0.8821317076683044, 'val_auc_std', 0.020633453907999904, 'val_accuracy_std', 0.02709243625153042)\n",
            "fract done 0.8541666666666666\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            3137        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 9,521\n",
            "Trainable params: 9,521\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.9399999976158142\n",
            "Number of epochs: 92\n",
            "Seconds per hyperparam config 25.04737424850464\n",
            "('mean_val_auc', 0.8712338089942933, 'mean_val_accuracy', 0.7911999940872192, 'metric', 0.8312169015407562, 'val_auc_std', 0.022122369245189486, 'val_accuracy_std', 0.02657368936976897)\n",
            "fract done 0.8611111111111112\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 384)          0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            12865       concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 19,249\n",
            "Trainable params: 19,249\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 86\n",
            "Seconds per hyperparam config 24.61746120452881\n",
            "('mean_val_auc', 0.8021878480911255, 'mean_val_accuracy', 0.7410000085830688, 'metric', 0.7715939283370972, 'val_auc_std', 0.03164349007358438, 'val_accuracy_std', 0.015257785348243803)\n",
            "fract done 0.8680555555555556\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            30753       sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 40,849\n",
            "Trainable params: 40,849\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.9200000166893005\n",
            "Number of epochs: 105\n",
            "Seconds per hyperparam config 25.430397033691406\n",
            "('mean_val_auc', 0.8868884325027466, 'mean_val_accuracy', 0.8192000150680542, 'metric', 0.8530442237854003, 'val_auc_std', 0.029839749297300225, 'val_accuracy_std', 0.03308413289184341)\n",
            "fract done 0.875\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 384)          0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            6177        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 12,561\n",
            "Trainable params: 12,561\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 92\n",
            "Seconds per hyperparam config 20.813225984573364\n",
            "('mean_val_auc', 0.7789158463478089, 'mean_val_accuracy', 0.7127999901771546, 'metric', 0.7458579182624817, 'val_auc_std', 0.05868890794518563, 'val_accuracy_std', 0.05858123044341634)\n",
            "fract done 0.8819444444444444\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 1984)         0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            31777       concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 32,657\n",
            "Trainable params: 32,657\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 115\n",
            "Seconds per hyperparam config 24.620139598846436\n",
            "('mean_val_auc', 0.810591971874237, 'mean_val_accuracy', 0.752400016784668, 'metric', 0.7814959943294525, 'val_auc_std', 0.05096191285454317, 'val_accuracy_std', 0.04935017540232907)\n",
            "fract done 0.8888888888888888\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            3137        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 4,017\n",
            "Trainable params: 4,017\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.7599999904632568\n",
            "Number of epochs: 173\n",
            "Seconds per hyperparam config 55.509870529174805\n",
            "('mean_val_auc', 0.8649060010910035, 'mean_val_accuracy', 0.7565999984741211, 'metric', 0.8107529997825622, 'val_auc_std', 0.02974014357241261, 'val_accuracy_std', 0.09599916352046782)\n",
            "fract done 0.8958333333333334\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            5153        sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 29,969\n",
            "Trainable params: 29,969\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 57\n",
            "Seconds per hyperparam config 21.74859642982483\n",
            "('mean_val_auc', 0.9074839234352112, 'mean_val_accuracy', 0.8553999900817871, 'metric', 0.8814419567584992, 'val_auc_std', 0.011117836032499431, 'val_accuracy_std', 0.013994278834158741)\n",
            "fract done 0.9027777777777778\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 1984)         0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            64065       concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 64,945\n",
            "Trainable params: 64,945\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 95\n",
            "Seconds per hyperparam config 28.34855628013611\n",
            "('mean_val_auc', 0.8474615812301636, 'mean_val_accuracy', 0.7905999898910523, 'metric', 0.8190307855606079, 'val_auc_std', 0.032620445249709994, 'val_accuracy_std', 0.03871743461782333)\n",
            "fract done 0.9097222222222222\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1313        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 2,193\n",
            "Trainable params: 2,193\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 75\n",
            "Seconds per hyperparam config 22.676007747650146\n",
            "('mean_val_auc', 0.885722541809082, 'mean_val_accuracy', 0.8213999986648559, 'metric', 0.853561270236969, 'val_auc_std', 0.021600435561111186, 'val_accuracy_std', 0.029323040108911486)\n",
            "fract done 0.9166666666666666\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1089        sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 25,905\n",
            "Trainable params: 25,905\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.9800000190734863\n",
            "Number of epochs: 93\n",
            "Seconds per hyperparam config 27.300094604492188\n",
            "('mean_val_auc', 0.8955438613891602, 'mean_val_accuracy', 0.8314000010490418, 'metric', 0.863471931219101, 'val_auc_std', 0.017332146816668663, 'val_accuracy_std', 0.03430801676327383)\n",
            "fract done 0.9236111111111112\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            62017       sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 72,113\n",
            "Trainable params: 72,113\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.9599999785423279\n",
            "Number of epochs: 61\n",
            "Seconds per hyperparam config 29.12208342552185\n",
            "('mean_val_auc', 0.8717908143997193, 'mean_val_accuracy', 0.8156000018119812, 'metric', 0.8436954081058503, 'val_auc_std', 0.02137269548540253, 'val_accuracy_std', 0.023035622674357326)\n",
            "fract done 0.9305555555555556\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.1, 'drop_rate': 0.4, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1089        sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 25,905\n",
            "Trainable params: 25,905\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 104\n",
            "Seconds per hyperparam config 23.330012798309326\n",
            "('mean_val_auc', 0.9092166304588318, 'mean_val_accuracy', 0.8531999945640564, 'metric', 0.881208312511444, 'val_auc_std', 0.012968965233138882, 'val_accuracy_std', 0.032102344828381396)\n",
            "fract done 0.9375\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            10817       sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 35,633\n",
            "Trainable params: 35,633\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.9800000190734863\n",
            "Number of epochs: 103\n",
            "Seconds per hyperparam config 33.00578689575195\n",
            "('mean_val_auc', 0.9001694083213806, 'mean_val_accuracy', 0.8450000047683716, 'metric', 0.8725847065448761, 'val_auc_std', 0.021829149022692763, 'val_accuracy_std', 0.02434747603650304)\n",
            "fract done 0.9444444444444444\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1089        sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 25,905\n",
            "Trainable params: 25,905\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 102\n",
            "Seconds per hyperparam config 27.95959234237671\n",
            "('mean_val_auc', 0.8942238688468933, 'mean_val_accuracy', 0.8269999980926513, 'metric', 0.8606119334697724, 'val_auc_std', 0.015681753763103636, 'val_accuracy_std', 0.02379075085892611)\n",
            "fract done 0.9513888888888888\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 80)           0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1313        concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 7,697\n",
            "Trainable params: 7,697\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 62\n",
            "Seconds per hyperparam config 18.623215436935425\n",
            "('mean_val_auc', 0.8272026062011719, 'mean_val_accuracy', 0.7431999921798706, 'metric', 0.7852012991905213, 'val_auc_std', 0.031308238320328384, 'val_accuracy_std', 0.03440000712871745)\n",
            "fract done 0.9583333333333334\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 0.5, 'drop_rate': 0.0, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           24816       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            289         sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 25,105\n",
            "Trainable params: 25,105\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 82\n",
            "Seconds per hyperparam config 24.733319759368896\n",
            "('mean_val_auc', 0.9069627165794373, 'mean_val_accuracy', 0.849399995803833, 'metric', 0.8781813561916352, 'val_auc_std', 0.01533445292329608, 'val_accuracy_std', 0.02722204808046835)\n",
            "fract done 0.9652777777777778\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            30753       sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 40,849\n",
            "Trainable params: 40,849\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 71\n",
            "Seconds per hyperparam config 22.291451930999756\n",
            "('mean_val_auc', 0.8769471883773804, 'mean_val_accuracy', 0.7973999857902527, 'metric', 0.8371735870838166, 'val_auc_std', 0.018273276132553758, 'val_accuracy_std', 0.09074492412148964)\n",
            "fract done 0.9722222222222222\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((32, 3, 1), (16, 3, 2)), 'reg_amount': 1, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 320)          6384        img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 384)          0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            12865       concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 19,249\n",
            "Trainable params: 19,249\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 83\n",
            "Seconds per hyperparam config 25.96096420288086\n",
            "('mean_val_auc', 0.8035986185073852, 'mean_val_accuracy', 0.746999990940094, 'metric', 0.7752993047237395, 'val_auc_std', 0.040478370407192786, 'val_accuracy_std', 0.04782467329648911)\n",
            "fract done 0.9791666666666666\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            62017       sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 72,113\n",
            "Trainable params: 72,113\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.9800000190734863\n",
            "Number of epochs: 137\n",
            "Seconds per hyperparam config 32.43969249725342\n",
            "('mean_val_auc', 0.9004592418670654, 'mean_val_accuracy', 0.831000006198883, 'metric', 0.8657296240329742, 'val_auc_std', 0.023521544933868762, 'val_accuracy_std', 0.04609121459815936)\n",
            "fract done 0.9861111111111112\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1089        sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 11,185\n",
            "Trainable params: 11,185\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 69\n",
            "Seconds per hyperparam config 20.865665674209595\n",
            "('mean_val_auc', 0.9001528978347778, 'mean_val_accuracy', 0.8375999927520752, 'metric', 0.8688764452934266, 'val_auc_std', 0.0030280952075779377, 'val_accuracy_std', 0.007964935201055552)\n",
            "fract done 0.9930555555555556\n",
            "==============================================================================================\n",
            "hparams {'concat_top_layer': 1, 'do_ave_pool': 0, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 1, 'drop_rate': 0.0, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 1920)         880         img_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 1984)         0           sequential[0][0]                 \n",
            "                                                                 dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            64065       concatenate[0][0]                \n",
            "==================================================================================================\n",
            "Total params: 64,945\n",
            "Trainable params: 64,945\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 1.0\n",
            "Number of epochs: 99\n",
            "Seconds per hyperparam config 35.731210708618164\n",
            "('mean_val_auc', 0.8342516660690308, 'mean_val_accuracy', 0.7658000111579895, 'metric', 0.8000258386135102, 'val_auc_std', 0.027697723686042203, 'val_accuracy_std', 0.03166322895089042)\n",
            "fract done 1.0\n",
            "==============================================================================================\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YHbqZtL4c_v9",
        "colab_type": "code",
        "outputId": "91347657-1c29-4140-ed69-dbf5730f7b56",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 547
        }
      },
      "source": [
        "print(\"best_hparams\", best_hparams)\n",
        "print(\"best results\", best_print_data)\n",
        "print(\"Retraining on the best_hparams to make sure we didn't just get good results by random chance.\")\n",
        "\n",
        "_, print_data = multiple_train_ave(best_hparams)\n",
        "print(\"Result of retrain on the best hyperparameters\", print_data)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "best_hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
            "best results ('mean_val_auc', 0.9142109990119934, 'mean_val_accuracy', 0.8658000111579895, 'metric', 0.8900055050849915, 'val_auc_std', 0.011914377924630034, 'val_accuracy_std', 0.014743140828623479)\n",
            "Retraining on the best_hparams to make sure we didn't just get good results by random chance.\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "dense_input (InputLayer)        [(None, 64)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "reshape (Reshape)               (None, 1, 1, 64)     0           dense_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "img_input (InputLayer)          [(None, 12, 14, 6)]  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "tf_op_layer_Tile (TensorFlowOpL [(None, 12, 14, 64)] 0           reshape[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "concatenate (Concatenate)       (None, 12, 14, 70)   0           img_input[0][0]                  \n",
            "                                                                 tf_op_layer_Tile[0][0]           \n",
            "__________________________________________________________________________________________________\n",
            "sequential (Sequential)         (None, 16)           10096       concatenate[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "output (Sequential)             (None, 1)            1089        sequential[0][0]                 \n",
            "==================================================================================================\n",
            "Total params: 11,185\n",
            "Trainable params: 11,185\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "best train accuracy: 0.9599999785423279\n",
            "Number of epochs: 122\n",
            "Seconds per hyperparam config 32.107903242111206\n",
            "Result of retrain on the best hyperparameters ('mean_val_auc', 0.9018999218940735, 'mean_val_accuracy', 0.8467999935150147, 'metric', 0.8743499577045442, 'val_auc_std', 0.011060894240953769, 'val_accuracy_std', 0.014496891143596702)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5FNKIjaxhw_g",
        "colab_type": "text"
      },
      "source": [
        "Results:\n",
        "\n",
        "on conv activations:\n",
        "```\n",
        "best_hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.4, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
        "best results ('mean_val_auc', 0.9149016737937927, 'mean_val_accuracy', 0.8551999926567078, 'metric', 0.8850508332252502, 'val_auc_std', 0.002876704858882324, 'val_accuracy_std', 0.011142719509147752)\n",
        "Result of retrain on the best hyperparameters ('mean_val_auc', 0.8857080936431885, 'mean_val_accuracy', 0.8265999913215637, 'metric', 0.8561540424823761, 'val_auc_std', 0.03448454328622984, 'val_accuracy_std', 0.027782011232540788)\n",
        "```\n",
        "\n",
        "on reduced_train_attri_opt_act and dense:\n",
        "```\n",
        "best_hparams {'concat_top_layer': 0, 'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.5, 'drop_rate': 0.2, 'fc_layer_sizes': (32, 16), 'reduction_alg': None, 'n_components': None}\n",
        "best results ('mean_val_auc', 0.9142109990119934, 'mean_val_accuracy', 0.8658000111579895, 'metric', 0.8900055050849915, 'val_auc_std', 0.011914377924630034, 'val_accuracy_std', 0.014743140828623479)\n",
        "Result of retrain on the best hyperparameters ('mean_val_auc', 0.9018999218940735, 'mean_val_accuracy', 0.8467999935150147, 'metric', 0.8743499577045442, 'val_auc_std', 0.011060894240953769, 'val_accuracy_std', 0.014496891143596702)\n",
        "```\n",
        "\n",
        "on reduced_train_act_opt_act:\n",
        "\n",
        "```\n",
        "best_hparams {'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.0, 'drop_rate': 0.2, 'fc_layer_sizes': (), 'reduction_alg': None, 'n_components': None}\n",
        "best results ('mean_val_auc', 0.8338292956352233, 'mean_val_accuracy', 0.770799994468689, 'metric', 0.8023146450519562, 'val_auc_std', 0.020978952292563628, 'val_accuracy_std', 0.02503918087631713)\n",
        "Result of retrain on the best hyperparameters ('mean_val_auc', 0.8178006529808044, 'mean_val_accuracy', 0.7424000144004822, 'metric', 0.7801003336906434, 'val_auc_std', 0.04277533943436206, 'val_accuracy_std', 0.0313343298064392)\n",
        "```\n",
        "\n",
        "on reduced_train_attri_opt_act:\n",
        "\n",
        "```\n",
        "best_hparams {'do_ave_pool': 1, 'conv_layer_params': ((16, 3, 1),), 'reg_amount': 0.1, 'drop_rate': 0.2, 'fc_layer_sizes': (16,), 'reduction_alg': None, 'n_components': None}\n",
        "best results ('mean_val_auc', 0.8876911640167237, 'mean_val_accuracy', 0.8391999959945678, 'metric', 0.8634455800056458, 'val_auc_std', 0.01197209366602662, 'val_accuracy_std', 0.017151083603375163)\n",
        "Result of retrain on the best hyperparameters ('mean_val_auc', 0.8815893292427063, 'mean_val_accuracy', 0.8297999978065491, 'metric', 0.8556946635246276, 'val_auc_std', 0.017778033902734827, 'val_accuracy_std', 0.023429907808406807)\n",
        "```"
      ]
    }
  ]
}