02_image_classification_fashionmnist.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "from tensorflow import keras\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import os\n",
    "import time"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "fashion_mnist = keras.datasets.fashion_mnist\n",
    "(train_images, train_labels), (test_images, test_labels) = fashion_mnist.load_data()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000, 28, 28)"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_images.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(60000,)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_labels.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "9"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_labels[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_images = train_images / 255.0\n",
    "test_images = test_images / 255.0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "class_names = [\"T-shirt/top\", \"Trouser\", \"Pullover\", \"Dress\", \"Coat\", \"Sandal\", \"Shirt\", \"Sneaker\", \"Bag\", \"Ankle boot\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 20 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(10,10))\n",
    "for i in range(20):\n",
    "    plt.subplot(5,4, i+1)\n",
    "    plt.xticks([])\n",
    "    plt.imshow(train_images[i])\n",
    "    plt.xlabel(class_names[train_labels[i]])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "validation_images = train_images[:5000]\n",
    "validation_labels = train_labels[:5000]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'T-shirt/top'"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "class_names[train_labels[2]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Classification MLP(Multilayer perceptron) with two hidden layers\n",
    "model = keras.models.Sequential([\n",
    "    keras.layers.Flatten(input_shape=[28,28]),\n",
    "    keras.layers.Dense(500, activation=keras.activations.relu),\n",
    "    keras.layers.Dense(250, activation=keras.activations.relu),\n",
    "    keras.layers.Dense(100, activation=keras.activations.relu),\n",
    "    keras.layers.Dense(10, activation=keras.activations.softmax)\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "flatten (Flatten)            (None, 784)               0         \n",
      "_________________________________________________________________\n",
      "dense (Dense)                (None, 500)               392500    \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 250)               125250    \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 100)               25100     \n",
      "_________________________________________________________________\n",
      "dense_3 (Dense)              (None, 10)                1010      \n",
      "=================================================================\n",
      "Total params: 543,860\n",
      "Trainable params: 543,860\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<tf.Variable 'dense/kernel:0' shape=(784, 500) dtype=float32, numpy=\n",
      "array([[ 0.05633341,  0.06464744,  0.0604151 , ...,  0.04696436,\n",
      "         0.00849533, -0.0564543 ],\n",
      "       [-0.04869322, -0.03663551,  0.00244601, ..., -0.05907154,\n",
      "         0.03748684, -0.00877194],\n",
      "       [-0.00634693, -0.0175994 , -0.02692286, ..., -0.03518824,\n",
      "        -0.01544062, -0.00769622],\n",
      "       ...,\n",
      "       [-0.04919377,  0.01352403, -0.03093696, ..., -0.02598281,\n",
      "        -0.02880395,  0.06334034],\n",
      "       [ 0.01062202,  0.0382546 , -0.0438368 , ..., -0.00061513,\n",
      "        -0.01323135,  0.02523389],\n",
      "       [ 0.05532134, -0.04823146, -0.03316089, ..., -0.01920315,\n",
      "         0.03375428,  0.03071746]], dtype=float32)>\n",
      "_____________\n",
      "_____________\n",
      "<tf.Variable 'dense/bias:0' shape=(500,) dtype=float32, numpy=\n",
      "array([0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
      "       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
      "       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
      "       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
      "       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
      "       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
      "       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
      "       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
      "       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
      "       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
      "       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
      "       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
      "       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
      "       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
      "       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
      "       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
      "       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
      "       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
      "       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
      "       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
      "       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
      "       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
      "       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
      "       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
      "       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
      "       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
      "       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
      "       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
      "       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
      "       0., 0., 0., 0., 0., 0., 0.], dtype=float32)>\n"
     ]
    }
   ],
   "source": [
    "first_hidden_layer = model.layers[1]\n",
    "weights, biases = first_hidden_layer.weights\n",
    "print(weights)\n",
    "print('_____________')\n",
    "print('_____________')\n",
    "print(biases)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "sgd = keras.optimizers.SGD(lr=0.01, decay=1e-6, momentum=0.9, nesterov=True)\n",
    "model.compile(loss=\"sparse_categorical_crossentropy\", optimizer=sgd, metrics=[\"accuracy\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "root_logdir = os.path.join(os.curdir, \"runs\")\n",
    "def get_run_logdir():\n",
    "    run_id = time.strftime(\"run_%Y_%m_%d-%H_%M_%S\")\n",
    "    return os.path.join(root_logdir, run_id)\n",
    "run_logdir = get_run_logdir()\n",
    "tensorboard_cb = keras.callbacks.TensorBoard(run_logdir)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 60000 samples, validate on 5000 samples\n",
      "Epoch 1/60\n",
      "60000/60000 [==============================] - 9s 144us/sample - loss: 0.4886 - accuracy: 0.8246 - val_loss: 0.3527 - val_accuracy: 0.8732\n",
      "Epoch 2/60\n",
      "60000/60000 [==============================] - 8s 129us/sample - loss: 0.3649 - accuracy: 0.8654 - val_loss: 0.3079 - val_accuracy: 0.8868\n",
      "Epoch 3/60\n",
      "60000/60000 [==============================] - 8s 133us/sample - loss: 0.3279 - accuracy: 0.8784 - val_loss: 0.2960 - val_accuracy: 0.8906\n",
      "Epoch 4/60\n",
      "60000/60000 [==============================] - 8s 132us/sample - loss: 0.3022 - accuracy: 0.8861 - val_loss: 0.2670 - val_accuracy: 0.8994\n",
      "Epoch 5/60\n",
      "60000/60000 [==============================] - 8s 134us/sample - loss: 0.2860 - accuracy: 0.8921 - val_loss: 0.2479 - val_accuracy: 0.9100\n",
      "Epoch 6/60\n",
      "60000/60000 [==============================] - 10s 159us/sample - loss: 0.2699 - accuracy: 0.8976 - val_loss: 0.2344 - val_accuracy: 0.9144\n",
      "Epoch 7/60\n",
      "60000/60000 [==============================] - 10s 161us/sample - loss: 0.2573 - accuracy: 0.9031 - val_loss: 0.2243 - val_accuracy: 0.9132\n",
      "Epoch 8/60\n",
      "60000/60000 [==============================] - 10s 172us/sample - loss: 0.2440 - accuracy: 0.9074 - val_loss: 0.2444 - val_accuracy: 0.9080\n",
      "Epoch 9/60\n",
      "60000/60000 [==============================] - 9s 156us/sample - loss: 0.2370 - accuracy: 0.9094 - val_loss: 0.2101 - val_accuracy: 0.9164\n",
      "Epoch 10/60\n",
      "60000/60000 [==============================] - 9s 152us/sample - loss: 0.2263 - accuracy: 0.9147 - val_loss: 0.2019 - val_accuracy: 0.9244\n",
      "Epoch 11/60\n",
      "60000/60000 [==============================] - 9s 148us/sample - loss: 0.2150 - accuracy: 0.9179 - val_loss: 0.2015 - val_accuracy: 0.9208\n",
      "Epoch 12/60\n",
      "60000/60000 [==============================] - 9s 143us/sample - loss: 0.2079 - accuracy: 0.9205 - val_loss: 0.1944 - val_accuracy: 0.9220\n",
      "Epoch 13/60\n",
      "60000/60000 [==============================] - 9s 149us/sample - loss: 0.2017 - accuracy: 0.9233 - val_loss: 0.1760 - val_accuracy: 0.9304\n",
      "Epoch 14/60\n",
      "60000/60000 [==============================] - 9s 144us/sample - loss: 0.1944 - accuracy: 0.9245 - val_loss: 0.1794 - val_accuracy: 0.9280\n",
      "Epoch 15/60\n",
      "60000/60000 [==============================] - 8s 141us/sample - loss: 0.1882 - accuracy: 0.9275 - val_loss: 0.1737 - val_accuracy: 0.9354\n",
      "Epoch 16/60\n",
      "60000/60000 [==============================] - 8s 137us/sample - loss: 0.1806 - accuracy: 0.9313 - val_loss: 0.1583 - val_accuracy: 0.9388\n",
      "Epoch 17/60\n",
      "60000/60000 [==============================] - 8s 136us/sample - loss: 0.1744 - accuracy: 0.9319 - val_loss: 0.1574 - val_accuracy: 0.9402\n",
      "Epoch 18/60\n",
      "60000/60000 [==============================] - 9s 143us/sample - loss: 0.1685 - accuracy: 0.9363 - val_loss: 0.1524 - val_accuracy: 0.9432\n",
      "Epoch 19/60\n",
      "60000/60000 [==============================] - 8s 141us/sample - loss: 0.1630 - accuracy: 0.9374 - val_loss: 0.1431 - val_accuracy: 0.9428\n",
      "Epoch 20/60\n",
      "60000/60000 [==============================] - 8s 136us/sample - loss: 0.1592 - accuracy: 0.9392 - val_loss: 0.1269 - val_accuracy: 0.9514\n",
      "Epoch 21/60\n",
      "60000/60000 [==============================] - 9s 153us/sample - loss: 0.1550 - accuracy: 0.9398 - val_loss: 0.1495 - val_accuracy: 0.9438\n",
      "Epoch 22/60\n",
      "60000/60000 [==============================] - 8s 137us/sample - loss: 0.1491 - accuracy: 0.9432 - val_loss: 0.1316 - val_accuracy: 0.9522\n",
      "Epoch 23/60\n",
      "60000/60000 [==============================] - 8s 139us/sample - loss: 0.1432 - accuracy: 0.9446 - val_loss: 0.1310 - val_accuracy: 0.9492\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.keras.callbacks.History at 0x16940489388>"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "early_stopping_cb = keras.callbacks.EarlyStopping(monitor='val_loss', min_delta=0, patience=3, verbose=0, mode='auto')\n",
    "model.fit(train_images, train_labels, epochs=60, validation_data=(validation_images, validation_labels), callbacks=[tensorboard_cb, early_stopping_cb])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "10000/10000 [==============================] - 0s 40us/sample - loss: 0.3596 - accuracy: 0.8907\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[0.3596461825724691, 0.8907]"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.evaluate(test_images, test_labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[9.90452945e-12, 7.05926160e-12, 3.03510683e-09, 2.89017907e-11,\n",
       "        1.51396361e-11, 3.62973722e-08, 1.70793268e-10, 6.79931645e-06,\n",
       "        2.73775197e-13, 9.99993205e-01],\n",
       "       [2.46195239e-04, 1.13196412e-08, 9.98627305e-01, 8.14729617e-08,\n",
       "        1.00012787e-03, 6.17060136e-09, 1.26182946e-04, 1.92714857e-08,\n",
       "        1.14798864e-08, 6.68559395e-08],\n",
       "       [1.05545933e-19, 1.00000000e+00, 9.06413013e-26, 1.13806050e-16,\n",
       "        1.97752208e-19, 1.04813020e-24, 1.06799301e-14, 3.55213553e-20,\n",
       "        1.45966010e-24, 2.45936832e-26],\n",
       "       [6.93134419e-16, 1.00000000e+00, 1.98763842e-22, 3.10907289e-12,\n",
       "        8.20796896e-17, 1.53123097e-21, 7.80685239e-12, 4.27689145e-18,\n",
       "        1.20801338e-21, 8.40224106e-22],\n",
       "       [6.40900731e-02, 6.10230000e-09, 2.34254403e-03, 1.37179927e-06,\n",
       "        6.01002066e-05, 1.94358769e-08, 9.33504879e-01, 2.14288363e-07,\n",
       "        2.81555739e-08, 8.53668439e-07],\n",
       "       [2.80717730e-14, 1.00000000e+00, 7.44165219e-20, 7.57839863e-13,\n",
       "        1.31721607e-15, 3.78118899e-18, 7.17455748e-11, 1.20348188e-15,\n",
       "        3.31104600e-18, 2.18121437e-20],\n",
       "       [1.45683433e-13, 3.32022177e-15, 2.38251872e-08, 2.32082452e-12,\n",
       "        9.99996424e-01, 4.33287627e-17, 3.59925366e-06, 1.26111203e-17,\n",
       "        1.62486051e-12, 6.81158739e-15],\n",
       "       [6.34631930e-12, 1.21003230e-15, 1.02406250e-06, 2.52869492e-12,\n",
       "        1.09707253e-04, 1.11010801e-16, 9.99889255e-01, 3.91895175e-14,\n",
       "        1.03448974e-16, 1.51848369e-13],\n",
       "       [4.13709955e-10, 2.00321201e-11, 3.06349564e-11, 2.79689726e-12,\n",
       "        2.89410558e-14, 1.00000000e+00, 9.42914434e-14, 2.76574630e-09,\n",
       "        3.00118402e-11, 1.42513666e-14],\n",
       "       [9.93949801e-12, 4.12396610e-11, 1.40366505e-12, 1.43892190e-11,\n",
       "        3.29868988e-12, 1.55769698e-07, 3.44037537e-12, 9.99998569e-01,\n",
       "        3.30751981e-10, 1.24532653e-06]], dtype=float32)"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "practical_test_images =  test_images[:10]\n",
    "prediction_probabilites = model.predict(practical_test_images)\n",
    "prediction_probabilites"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[9, 2, 1, 1, 6, 1, 4, 6, 5, 7]"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def derive_predicted_classes(prediction_probabilites):\n",
    "    batch_prediction = []\n",
    "    for vector in prediction_probabilites:\n",
    "        batch_prediction.append(np.argmax(vector))\n",
    "    return batch_prediction\n",
    "model_prediction = derive_predicted_classes(prediction_probabilites)\n",
    "model_prediction\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([9, 2, 1, 1, 6, 1, 4, 6, 5, 7], dtype=int64)"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model_prediction = model.predict_classes(practical_test_images)\n",
    "model_prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['Ankle boot', 'Pullover', 'Trouser', 'Trouser', 'Shirt', 'Trouser',\n",
       "       'Coat', 'Shirt', 'Sandal', 'Sneaker'], dtype='<U11')"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.array(class_names)[model_prediction]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x720 with 10 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Visualise the prediction result\n",
    "plt.figure(figsize=(10,10))\n",
    "for i in range(len(practical_test_images)):\n",
    "    plt.subplot(5,5, i+1)\n",
    "    plt.xticks([])\n",
    "    plt.yticks([])\n",
    "    plt.grid(False)\n",
    "    plt.imshow(practical_test_images[i])\n",
    "    plt.xlabel(class_names[model_prediction[i]])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.save(\"image_classification_model.h5\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[9 2 1 1 6 1 4 6 5 7]\n",
      "['Ankle boot' 'Pullover' 'Trouser' 'Trouser' 'Shirt' 'Trouser' 'Coat'\n",
      " 'Shirt' 'Sandal' 'Sneaker']\n"
     ]
    }
   ],
   "source": [
    "loaded_model = keras.models.load_model(\"image_classification_model.h5\")\n",
    "predictions = loaded_model.predict_classes(practical_test_images)\n",
    "print(predictions)\n",
    "print(np.array(class_names)[predictions])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}