4_1_ML_tutorial_chatGPT.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "colab_type": "text",
        "id": "view-in-github"
      },
      "source": [
        "[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/mlexchange/als_ml_tutorial/blob/main/4_1_ML_tutorial_chatGPT.ipynb)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UC6u4hELhfd_"
      },
      "source": [
        "# **chatGPT**\n",
        "by OpenAI\n",
        "What can we do with it?\n",
        "ALWAYS be aware the answer could be VERY wrong.\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {},
      "source": [
        "Before working with chatGPT, we need to load data"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {},
      "outputs": [],
      "source": [
        "# Download the data (400 scattering images 1.3Mb)\n",
        "# If not on Google Colab or on a machine without wget, comment out the two cells below,\n",
        "# download the linked file, and unzip in the directory of this notebook\n",
        "!wget -q 'https://drive.google.com/uc?id=1RTObsZOHLS4RUnwCegoxXD5EktvrSeI5' -O 'data_64x64_400.zip'"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 2,
      "metadata": {},
      "outputs": [],
      "source": [
        "!unzip -q -n 'data_64x64_400.zip' && rm 'data_64x64_400.zip'\n",
        "!mv \"data_64x64_400\" \"images\""
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "chuE7lGchemR"
      },
      "source": [
        "Let see what we can do\n",
        "\n",
        "\n",
        "1.   Read data\n",
        "2.   Show the data\n",
        "3.   Convert the data to a *tensor*\n",
        "4.   Build a simple autoencoder\n",
        "5.   Show input and output\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rqbwfi2Li4wI"
      },
      "source": [
        "Input to chatGPT 4.0:\n",
        "\n",
        "*\"I have a folder called \"images\" with jpg images and I want to load them and convert them to a pytorch tensor\"*"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "D1QcKSI0jS99"
      },
      "source": [
        "Output:"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "4qrf2bISjKdZ"
      },
      "outputs": [],
      "source": [
        "import os\n",
        "from PIL import Image\n",
        "import torch\n",
        "from torchvision import transforms\n",
        "\n",
        "# Initialize transformation\n",
        "transform = transforms.Compose([\n",
        "    transforms.Resize((224, 224)),  # Resize images to 224x224\n",
        "    transforms.ToTensor(),  # Convert PIL image to PyTorch tensor\n",
        "])\n",
        "\n",
        "# Folder path\n",
        "folder_path = 'images'\n",
        "\n",
        "# Read image file names from folder\n",
        "image_files = [f for f in os.listdir(folder_path) if f.endswith('.jpg')]\n",
        "\n",
        "# Initialize tensor for storing images\n",
        "tensor_list = []\n",
        "\n",
        "# Loop through image files to open, transform, and convert to tensor\n",
        "for image_file in image_files:\n",
        "    img_path = os.path.join(folder_path, image_file)\n",
        "    with Image.open(img_path) as img:\n",
        "        img_tensor = transform(img)\n",
        "        tensor_list.append(img_tensor)\n",
        "\n",
        "# Convert list of tensors to a single 4D tensor (Batch x Channels x Height x Width)\n",
        "image_tensor = torch.stack(tensor_list)\n",
        "\n",
        "# Print shape to confirm\n",
        "print(image_tensor.shape)\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Ca880C9ZjXZj"
      },
      "source": [
        "ahh forgot the size:\n",
        "\n",
        "\"my images are 128 by 128 pixels, don't resize\"\n",
        "(THIS IS NOT TRUE, MY IMAGES ARE 64x64)\n",
        "\n",
        "output:"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "IUm6EdeQjqyM"
      },
      "outputs": [],
      "source": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "yA3xuP1cjkkY",
        "outputId": "b7e87307-5db6-48d4-c169-c5d7e8fa9bb6"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "torch.Size([400, 1, 64, 64])\n"
          ]
        }
      ],
      "source": [
        "import os\n",
        "from PIL import Image\n",
        "import torch\n",
        "from torchvision import transforms\n",
        "\n",
        "# Initialize transformation\n",
        "transform = transforms.Compose([\n",
        "    transforms.ToTensor(),  # Convert PIL image to PyTorch tensor\n",
        "])\n",
        "\n",
        "# Folder path\n",
        "folder_path = 'images'\n",
        "\n",
        "# Read image file names from folder\n",
        "image_files = [f for f in os.listdir(folder_path) if f.endswith('.jpg')]\n",
        "\n",
        "# Initialize tensor for storing images\n",
        "tensor_list = []\n",
        "\n",
        "# Loop through image files to open, transform, and convert to tensor\n",
        "for image_file in image_files:\n",
        "    img_path = os.path.join(folder_path, image_file)\n",
        "    with Image.open(img_path) as img:\n",
        "        img_tensor = transform(img)\n",
        "        tensor_list.append(img_tensor)\n",
        "\n",
        "# Convert list of tensors to a single 4D tensor (Batch x Channels x Height x Width)\n",
        "image_tensor = torch.stack(tensor_list)\n",
        "\n",
        "# Print shape to confirm\n",
        "print(image_tensor.shape)\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rxvZUO3FkRn0"
      },
      "source": [
        "now I want to take a look\n",
        "\"show the last image using matplotlib\""
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 424
        },
        "id": "HKK24hDXkKqQ",
        "outputId": "71f1580c-ce15-4ed6-f9a1-d87b82cce19e"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Shape of the 4D tensor: torch.Size([400, 1, 64, 64])\n"
          ]
        },
        {
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAGFCAYAAAASI+9IAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAAAqaElEQVR4nO3d665eZbnG8adaNt3MTrvftxRaWlqoYGvYJYRIDKLGzx4DXzwCj8ID8BhM1ITE+EVFoOwUKnRP95RZoJ2gBQVZX5ZPXIzrX8a95pirK2v9fx/vjDnesZ1P3jzXez9Lvvjiiy+aJEmtta/d6gOQJP3v4aAgSeocFCRJnYOCJKlzUJAkdQ4KkqTOQUGS1DkoSJK6pWM3/OEPfxjrS5YsGV3/+te/Hre94447Yv3OO+8cvf3XvpbHt88++yzW6Td7dD6ff/55rCf//Oc/S/V0LHR8tI+qf/zjH4PatWvX4rYbN26M9SmuFd23KVSPb4pjue2222I9Xe/WprufFXNzc7G+Zs2aQe3222+P29K1pWuY6lPde9pP+n9D/w+oTveNnqHKb4HpGlZ8+umnsU7H94tf/OIr9+k3BUlS56AgSeocFCRJnYOCJKlzUJAkdaPTR6SS4qEkw7Jly2Kdtk+fSekBmoWnOiUCUkqEPnMK1WQGbV85n0qKozW+97R95Zyq6bCKqVIv6RiryabK+VS73FevVTr2SpKsqvqcTHHfaB/0v2bp0vwvku5FSitRQoiSZ5VkFyU3F/K/yW8KkqTOQUGS1DkoSJI6BwVJUuegIEnqRqePaKacZueXL18+qFEvo2qPmjTDT2kAqk/Ru4RU+8JU9kHXu/qZ6TwpDUL1qinSI1Okj0g1ObSYfYsq/bBI9RlK9cU892pChs6n8nxOkYxrja9h6n1F/bAolVRNUiYLeWf9piBJ6hwUJEmdg4IkqXNQkCR1oyeaaSKGJlzS9jRRUp20ShMxU0wQt1abyKz+HH+KiebqRGvlZ/o0OVVpN9Ja7Rhp2ynuZ2WSvbX65Fw69mrgoXKe1X1Un6H0rFTbPFQWkiLVRaoq7SKqYQKaJKbJ4HS9qs8Vta5IE9O0CJBtLiRJk3BQkCR1DgqSpM5BQZLUOShIkrrR6aPqDPrf//730dvSvqdIAlVVjqV63JWWAdWf41eTJmk/9JmUwCCV9FX1WlW2nyIJU91PNSFTSfFU91FdNKmSPqJjqSS+FnMhJdp/9RpSWqmSyppiMZ3Wai00Pvnkk1gfw28KkqTOQUGS1DkoSJI6BwVJUuegIEnqRqePaHaeemxMkQaq9FWaIq1S3c8UPY5ay+dZXVCksnBKa62tXLly9Lb33ntvrFOPlrTAUmu5hxI9V9RvidIWlfQR9Yuha06LoaT90/tA+/j4449jPaX3Pvroo9K+6Twr71U1fVTpz1RNMFUX/JkifUT1Ss+hapKQrnklAUoLmo3hNwVJUuegIEnqHBQkSZ2DgiSpc1CQJHULTh9Vtq+u7DVF+qiaSqrM/FfTA5ScWbFixaA2Ozsbt6X6qlWrYn1mZibWU0ri3LlzcdvHH3989D5aq62mVn2uKPWR9lO999TnptKLh85nihQPpYyofuPGjVj//e9/H+vpWaHzqaSmWqv1cqLnqtpXKt3P6opxlX2T6sqS1dXhEnrexvCbgiSpc1CQJHUOCpKkzkFBktSNno04efJkacdpYrY6uVuZKJxqYQ46ljQZXl2UZtmyZbGeJvioVUT1J/M0yZUW4Xjvvffiti+//HKs04TYVIvbVPaRPrPa4oRUJv4Wc9Gg6r2n7c+fPx/rKaxA+75+/Xqs0+R2uoaV9iGt1QIMN9tPZdspnuWpFnWq/E9dSJshvylIkjoHBUlS56AgSeocFCRJnYOCJKkbnT7asGFDrFcSEZTKocVapmg5QcdH7R927NgR61u2bBnU0kI1rfF5UooltQagdg7URuBvf/tb6TNT8mPNmjVx2wMHDsQ6XXNKWaUEFx0fPROVREk19UFJjkqrA2rRQG0hKou1VNsf0DWkBVj27ds3qK1evTpuS+9VpYXI+++/H7elOqXj6NrSvUiqi+xUUkl036qfOcVCRWP4TUGS1DkoSJI6BwVJUuegIEnqHBQkSd3o9BGldSoLzVDqgeq0+E5K91A6atOmTbFOKSPaTzofSjdUE0J//etfBzVKFVAiixI/3/jGN2I9Hfvp06fjtk899VSsVxZUaa2WBqouhFNZwIfqlKghle2rSahKryB6rug+XLt2LdZTmo4SdtV3Nm2/Z8+euC0d90cffRTrlEo6c+bMoEbJpmofr8oiO7QtPT/0fyXd/yn6Pn2Z3xQkSZ2DgiSpc1CQJHUOCpKkzkFBktSNTh9VVgNqLaeSKKlEvYLWrVsX69u3bx/Udu7cGbdNq0m1xudDK0elOm1LiQBaTS0lnui4KfVRTTJ8+OGHg1pKQbXW2uXLl2OdUi9pVbfWclqJEjWU2KB6SitVV/Cia0j11Fuo0vdpqn1Twozeq9THq7XWdu3aNajRcVOCidI96Z2g81mxYkWs0/tzzz33xPr69esHtbNnz8ZtL1y4EOvpPWltmhUA6Vmmdzb9/6ympsbwm4IkqXNQkCR1DgqSpM5BQZLUOShIkrrR6aNKyoi2pz4qu3fvjvW0ElRrra1du3ZQoz48H3zwQaxTcobSFin5QH2SKDVFaZCUNqj2f6HVpygl8u677w5qb731Vty2mjKqrCY2RU+g1morTVEahI6F9p32Q+8J9ayi1dHS80aJNKqn96Q1Ts2l55NW40vJntZq/b0o2XP16tXRx9ca/19J13Dv3r1x261bt8b6xYsXY/3cuXOxnt43ShPR80bnmf4n0D7oeRvDbwqSpM5BQZLUOShIkjoHBUlS99+fjfhPNGmXJqgOHz4ct73rrrtKnzk3Nzeo0YQq2bhxY6zThNPs7OygRudOE3k0sZbqNNlG+6AJaJqATxN/tFgJTeTRBFpFddKX6pVFT6qfSdunenXivKKyoFVr3BaCWjqkkAUtRkUT0NSGJU1602Q1PcuV0ERreRKfWmhQncIutHhXaqNBrTXo/0TlWa6EV8bym4IkqXNQkCR1DgqSpM5BQZLUOShIkrrR6SNKPmzbti3WDx06NKhVfxp//fr1WE9tFChNRAuK0MIkdJ4p+UDHR+kOSvfMz88PatQqYiGLZ/y7lHCoJnuqaZ20H7reU5xnpfXFzT6T9lNJGlXPZ4oFVegZotYvb7/99qBWbTlBqZzNmzcPavQOUhIo7aO11mZmZmI9LfhD7ywtMEX7Xr16daync6J2OCdPnoz1K1euxDolCRNqtzKG3xQkSZ2DgiSpc1CQJHUOCpKkzkFBktSNTh8dOHAg1u+///5YX7Zs2aBGqQfq00Gz9qkfCyUZaCGYlExojY8xLbZByQzqaVLp51NNzlQWgmktX3NKLFA/m+qxpDodX/X8036qPY4IpXvSfuhZria4Up22peOj+0nXJaFUDklJutZaO3/+/KBGi1FR/zFKGNK7n/ZPCR5K/ND/CbrmKa20a9euuC0lI995551YT4tgUZ8oSp6N4TcFSVLnoCBJ6hwUJEmdg4IkqXNQkCR1o9NHZ86cifVz587FeupnRDPi1BOJ+qgklDKi2XlawYz6MKW0wVSrbE2xn2qKJ6VhKGlC16qSMmotp2EqSZjW+Fqlz5yqT1Sll1Nl1aybbT/Fvula0f1M7zL1OKr2vapsS59JKSOqpwQkJekoNfbxxx/HOqWs0up16Tha43QYHUvqnUb3kv4fjuE3BUlS56AgSeocFCRJnYOCJKkbPdFMPzFPi8+01tru3bsHtYceeihuSxNL1HLi8uXLgxotYEMTzTTJQ601pphAI+lY6JpQnSazli7Nt/jTTz8d1N5999247c6dO2OdggOV1g3/m0zRFqM64U/bp/pUiwadOnUq1lOLhsoke2vTXKsqmlRN7wotjrN9+/ZYTxPHrfG9SG0xqB0OvZu0sE86lvQet9ba8ePHY30MvylIkjoHBUlS56AgSeocFCRJnYOCJKkbnT6i1Mvhw4dj/dFHHx3UaCafFriYm5uL9ZR6oRn75cuXxzpZuXJlrKe0ASWY7rjjjlinhFA6RkomTJXsSS0taB979uyJdUpg0H1O50/PVfXapmOpXitKU9HCLKlOaRCqU0sDqieU4qFniM4zpcxoH9QOhpJA6XyqCaYpkl3VfVPybu3atbGerhddE0qB0TGmdh70/tDiZ2P4TUGS1DkoSJI6BwVJUuegIEnqHBQkSd3o9FHqZdRaa4cOHYr1lHA4evRo3JZm4amvUtp3dUGVah+ZlBCiNBElmChVkI79xo0bcVtK5VDih9JXq1atGtQoafLII4/EejVllbanzyR0/pWkSRWlZFKihlJGtIBRZREXeh9owRdKTVGyKR07PVfp+WmttRUrVsR6ep4pwUTXkO4DpanSO06L0tB9SH3WWmtt7969sZ7SSvT/YNeuXbF+/vz5WE+9ydasWRO3pfs2ht8UJEmdg4IkqXNQkCR1DgqSpM5BQZLUjY5+0KppaaWh1lp7++23B7V33nknbktJG+pHUkkaUcrozjvvjHWazU99R+g4KD1BaZiUEFq/fn3cdt26dbFOvVhoPynJ8eabb8Zt77nnnliv9u1J16uybWu1RBFtO8UqaK3l5BSlPmZnZ0v7TsdOKRtaXZBSSfTOplXJ6DOpnw+lyVICh5JxlEqihFDlfav2W6IVHelY0r2g5CY9E/S+Xbx4cVCjdBStMDeG3xQkSZ2DgiSpc1CQJHUOCpKkzkFBktSNTh9Rf6KUMmqttUuXLg1qlDShfjaVlAj14aFeLNS7hVJJKW1BCRnax6ZNm2I99UDZvHlz3JYSG5T6oGNMq9pRfx5aAY9SIpRYSfe/ujpaJSFU7X1USQLR9nQfKJVEK2el/lH0jFNijp63M2fOxHpKydAzQUnCq1evxnp6f+jc6d2knlqUvqJ6Qu8J1Sl99Je//GX0cdDqaJQYTP8T6F1LSaWx/KYgSeocFCRJnYOCJKlzUJAkdaMnml944YVYp5/MpwmaansBmoBOk8f0k3Ga4KPJQ1qYJE3w0kQe/ax927ZtsZ7Oh46P2gtQSwOaDL5y5cqgRpOEdI8JTbSna0iTjXTvaSK3smhStYUGPRNpko9aLtB9oPBFQudOYYrUmqU1Ps/UGoEWgtmxY0es0wRnCqpQCwm6VvSs0LuftqdJX3qvqhPQ6dgppEPHcvDgwVjfuHHjoEaBlOoiYv/lb//bfylJ+j/HQUGS1DkoSJI6BwVJUuegIEnqRqeP3n333dKOU3qEEiWUKqgseEMJJkqOUJKDUhUPPPDAoHbXXXfFbakdQSWxQMkEWpCIfu5O6at0bSnJcODAgVina67x6PlMaRhKmNG9v379eqx/8MEHo7enZBMlfuidTYk8apVx7NixWKd0HJmZmRnUqE0MXVtq80HXvIJSfa+99lqsp/eQ/gdR0nEMvylIkjoHBUlS56AgSeocFCRJnYOCJKlb8sXIlUgOHz6cd1BIoFDiJ6UEWmtt5cqVsZ5SPJ9//nnclk6P+sLs27cv1teuXRvrCSVKKDmU0keUVKKeJpSqoPRIul4XLlyI29I1qfYQSvVK759bhVJz6dmvJrIqiwbRM07XkJ63o0ePxnpa3KX6XFHyjlJwCaVyzp8/H+uVxXTo/aF7TPumd5zeiQo6lvT/kFJGtFDRT3/606/8fL8pSJI6BwVJUuegIEnqHBQkSZ2DgiSpG937qLpCVtp+3bp1cVtKOFB/kZRioTTRoUOHYn3Lli2xTkmOlBD68MMP47aUEqE01c6dOwc1SmRRgotSD3Q+qf8N9beq9pWiz0z1SlJpKrTvShKI0P2h94dSOWl7SpTQSnfLli2LdTr/9BzSc3Xt2rVYp5XnUuqH3tmtW7fG+oMPPhjr1Mvp7bffHr0t3WO65vTsp15J1WQkSfciraDYGj9vY/hNQZLUOShIkjoHBUlS56AgSeocFCRJ3ej00YYNG2Kd0kepj8rGjRvjttV+Idu3bx/UnnrqqbgtrUxEySbqu/Lee+8NapQqoPOka0jpkSSloFrjNAj1bkmpBVqp69577411Sk/Q+aRkDqVvqIcOPW8p3VJNNlFqipImqU7pG9oHSedJ15WSaukdbI3P84knnhjUKDFHzyGtjpZSP3Qv6T2htBIlbVKa7siRI3Hbs2fPxno1kZauLaUUK+lKQsdR6dX2ZX5TkCR1DgqSpM5BQZLUOShIkrrRE800KUI/A0+TXx9//PHYj2uttbZ///5Yf/rppwc1aqGRJohbyz+Bb43PM03a7dq1K247Ozsb6zSBlI6RjpsmrejaVlpOVCf+aOKTJomrC9BUTNEWY4oFcih8QBPNFARIdQpBUHsSWjTp5MmTsb5p06ZB7e67747bUmuatI/W8qI8tGjOiRMnYp0mT+n5XL169aD26KOPxm3p3tO1ouctTXrTZD29y/TOVp5xan8xht8UJEmdg4IkqXNQkCR1DgqSpM5BQZLUjU4fUdqAkjbpZ/CpFUFrrT3wwAOx/oMf/GD0Z16+fDlu++qrr8Y6pXUefvjhWE/tMujn9bRvSoOk9Ajtg1IS1AKAkhkpJUPJGUqY3bhxI9apXUZKdlVbnCymaoIpPc+VRXNa4/cn3U+6P/Pz87FOC8rQ9mfOnBnUqJ0FPVfU4iU9Q7t3747bXrp0KdYpUZMWtmmttW3bto06jtZae/zxx2Od3rdjx47FempzQvee3lm6bymVRO/PQt4rvylIkjoHBUlS56AgSeocFCRJnYOCJKkbnT6imfJPPvkk1lNSYu/evXHbZ555JtZp8ZDUM+Wll16K21IfkcceeyzWKRGR9kMpiZTiaI1TBSmVQ31etm7dGuvUX4WkJAcdH6Wm6N5Tn5+U7qmmJCr9iWjftI9q+ijth/pHUQKF+kSlFFPqH9Qav5vUD+ydd94ZfSzVxZuon096bun4aGGsZcuWxTqdz+nTpwe1HTt2xG3XrFkT65RKomcl9VSjfmr0TFBCKl3bKXp+fZnfFCRJnYOCJKlzUJAkdQ4KkqTOQUGS1I1OH1HfldTro7W8AtP3vve9uC3N/FOq4MUXXxzUKMXx1FNPxXrqi9Ia93pJx0LHR8kMSgilxBOljyjZQ8khOp+rV68OapQcmZubKx0LScmcalpn6dL8yKb9UDKDnmVKqtF5VvpHUeKpcv600h2lctLKY63xu3LvvfcOanTulEijFQNTfyLqnUUJuy1btsQ6XcNTp04NaimRdDOU7KIeaSmRl46jNU7HUf+s1H+O/tfY+0iSNAkHBUlS56AgSeocFCRJ3eiJZpoUosnTp59+elC777774rbnzp2L9SNHjoz+zPR5rfFEEU2qnjhxItbTBBr9fH3Xrl2xvnPnzlhPi7XQ8VHbAVqUh44xTfzR5BS1V6CAAE2IpjotvLSY7S+maGfRWr62levdGgcB0gQv7YMmg2ninJ6hNElKgQwKQtAEdPpMWhiL0PtDE9CpTc6bb74Zt6X/QTSJTy06nnzyyUGNwjhpca3W+J2oLLxELWjG8JuCJKlzUJAkdQ4KkqTOQUGS1DkoSJK6JV+MjGI88sgjsU4z/5QGSo4ePRrrlJL45je/OahRSoJm4ekzKWWVEiizs7NxWzoWSqakn6rTz9cpaUL7rqR7KN1y8ODBWKdHh44xJW0omUGpCvrMVKdt6ZpUW1GkNBWlVah1AZ1nuhfz8/NxW0oZ0XHT4lBpcRtKGVUXdUrpPXrG6VpR246NGzfGemqJQm04Ll68GOu00Bf930vtSc6ePRu3PXbsWKzTM5HOh941SiP+8pe/jPV/5zcFSVLnoCBJ6hwUJEmdg4IkqXNQkCR1o3sfUa+PJ554ItbTrPirr74at6XkDCWeUiLiypUrcdtXXnkl1imVlBYHaq21PXv2DGqUkqAFb6ifUUo80TWhRWaoDxHVU/Lj/fffj9vSYkLVHkJJpWfRVKp9lehepGeI7j2hJFTqN0XPJj0T169fj3W6nyntR0kYSresWrUq1jdv3jyoUXqPeiJRcoiOJaUA169fH7eld/nkyZOxTs/Ejh07BjXq2UTvz1tvvRXrKWVGCybRfRvDbwqSpM5BQZLUOShIkjoHBUlS56AgSepGp49+/OMfx/qPfvSjWH/hhRcGNUrfPPzww7FO/ZNSWufXv/513Jb6+VBi45lnnon1lCqglAStPkXpntT/h3roUJKBUh/0mbTqU/Ktb31r9LatcaIoJW1olTZa7Y22Tz1nCPVbolXQqB9WSthR6qOaeErbU9KEEjWUGHzuuediPT1DdB/oWGilw9SfiNJHc3NzsX7q1KlYpxRP6uWUUoStcZ8oWqntzJkzsX7gwIFBjXqhUe8n6k+UelZRamohyUC/KUiSOgcFSVLnoCBJ6hwUJEmdg4IkqRudPnrsscdi/fTp07H+hz/8YVCj1ZoeeuihWKc0yG9/+9tBjVZSozQInc+GDRtiPfWLoRWVqP8NrZKUVpTau3dv3JauyRtvvBHr1OOp0huF+vNQ4oeSUJV0y/Lly2Od0hbpWCiBQX1rqqmklGyjfkOUNKHPHPt5N9s3Jc9ISuvQs0znSeeT6tTLqbrCGiWBjh8/HuvJ3XffXapT8jClleiZpb5k999/f6yn/m6UjKuk8b7MbwqSpM5BQZLUOShIkjoHBUlSN3qimX4G/uKLL8Z6mvyiBXnop/F/+tOfYv31118f1GjyMP3s/GZ1Wsjj3LlzgxpN5NHkD02UpZ/eV46jNZ4MpnYZacEO2jYtkHKz7SlQkCaPqV0CPW90npXFeqjFCR0LTXrPzMwManTu9EzQhG2q0z6qC+Gkdgmt5XDDrl274rY0uTs/Px/rlTYfdH+oXQSdZ2rlQovm0H2jZ5+CIM8///ygdv78+bgtBTJ2794d62nxHQr6uMiOJGkSDgqSpM5BQZLUOShIkjoHBUlSNzp9dOLEiVg/duxYrKd2EbRYCy2+88c//jHWU9sBWmjk29/+dqxTAoPqKeFAaRD6mX5qI9BabpdBP6OnZAb9rJ0SNWmBE0pwUWqKEhv0mSkRQW04qEUFpY8orZSk5FVrnJCha55SSdX0ES0alPZDabdr167FOl0rkhJF1KJh+/btsZ7awbSWU0nV9il0H+i9SglI+l9DC/jQQkDUDiellagdDiWbaHGk/fv3D2qUbKJnfAy/KUiSOgcFSVLnoCBJ6hwUJEmdg4IkqRudPnrttddindIjDz744KBGyYzf/OY3sU5JoJROOHz4cNyWEjXUV4lSLymZQskRSkNcvHgx1i9cuDCoUeqDEhh0LJRwSNeQrhUlLajPDaVkUvqIznMxF9khlcV0WsuLJlFyhvp7bdmyJdbTtaWEFV0TSiVVklrp2Wytta1bt8b6zp07Yz2lF+n5oeOjZ5wW60nHSL2m6N2k94cW37nnnntG75uuLaUX07Wl1CWlksbwm4IkqXNQkCR1DgqSpM5BQZLUOShIkrrR6SNa8WvNmjWxnnoOUd+RN998M9apF01K91Dih1JGV65ciXVKSKVVku677764LaVVKE2VegVR0mLFihWxTr1oKD2R0j202ltawao17udDCalUTwmem+2bkimV9BF9Jj1v9JnpfGgfc3NzsU59olJyhlbqovtDK+PRZ6Znn86dPjOlb1rLK7hRPzVKNKZeRq1xsitdQ3rG6d2k/3uUskqrw+3YsSNuS8/Ep59+Guvpfy2tAEf3Zwy/KUiSOgcFSVLnoCBJ6hwUJEnd6InmkydPxjpNuDz//PODGi3IQ4vp0CIcaULsV7/6VdyWJrGvXr0a69ReIU1apYV3Wmvt9OnTsU4TnGlSmSZr165dG+s0aUVtFFILhJdffjluSz/TT20rWuPrkiaPaR9LliyJdZoknWKimdp80H7SJCw9PxQQoInp9OxTaICOjyZV6flMQRC63hQEeOONN2I9HTtNhlL7i5mZmVinsEt6V+i9TwsMtdba0aNHY50WwUptJ2gS+6233op1um+pnQcFZugej+E3BUlS56AgSeocFCRJnYOCJKlzUJAkdaPTR7QAybp162I9/YT9d7/7XdyWfr5OSZuUWHn11VfjtrSQBbWioLRO+kxaeIiSGbQAS/pZ/+zsbNy2mpChNh9pIRxKZlCCi9oRVNpcULqFUjyUSqq0nKB90LWltgMpxUTJps8++yzWK2klSpTQe1J9htI7Tu8m7YOSMx988MHo46PFgWjxJtq+kqaiZ7naiiK1+aDnrZrgSv+zKKFJ9TH8piBJ6hwUJEmdg4IkqXNQkCR1DgqSpG50+ujQoUOx/uyzz8Z6SidQrxOa+U8L9bTW2kMPPTSoHTlyJG5L/UVSGqK11r7zne/E+p49ewa1V155JW5LSRNaOGf16tWDWlqUpLXWDh48WPpMOv+UfHjppZfitvv37491WoCFFipK50kLpFCfG1o4poKSWpRIo15OKa1FSRjaNyWk0nWh/kmUaqNFoKhfWVrEh1KHlf5WreXUDx3f+fPnY53+f6R+Q63lhbcoLUk9jo4fPx7r1G/pu9/97qBGz3LqD9caJyaffPLJQS29U61x0nMMvylIkjoHBUlS56AgSeocFCRJnYOCJKkbnT6i2XZKRKTZb0r8UKKEZu3TSmC0AhGlQWjFL/rMc+fODWqUwKD+KpQQSttTPxvqo/L666/HOvVoST1drl+/HreltA6dJyUiUq8buvfV/kSUhEroPtA+qD8R3aOEjpt6C6XVuui6Uqrt1KlTsU69qdJ50rWi86H3KvUhoj5JKQXVWn4HW+P+ROldvv322+O29H+Mzof+36T7lpKLrfF50nuYPpPeH0r1jeE3BUlS56AgSeocFCRJnYOCJKlzUJAkdaPTR9u2bYt16ruS+nfQqlQ0C09JoMuXLw9q8/PzcVvqxUKz85SoSZ9JCRlK69C1SseycePGuG1KXrXGq6ZRoiYlHGi1JkrfUB8Zup8pKUFJGLqGVKd7UdkHqaR1KK1CPY4o3ZPqKcHTGvf+oWeC3pV9+/aNPj56VirvBKUR0+plrXHKit7xlA6kHlSUPqJ7T/tJq8PdfffdcVv6H0TXkFJJSSUZ92V+U5AkdQ4KkqTOQUGS1DkoSJK60RPNNKlIbRTSxCxNWtFPz2lCME240WQTLeBDk9h0jDSxVEGT2GnCia4JLcBB14omFdPEJ02E08QxTc7ReabJSQof0AQ57bsy0Uz7pmOha5v2U23ZQs9Vem4rE6qt8WQwTfCmxZGq95jOJ11DmjilyXoKPNB+0gJB9P+gGjyhz0zXlv6n0LWl5zO17KEAAx33GH5TkCR1DgqSpM5BQZLUOShIkjoHBUlSNzp9RLPzlDZICSFamIOSI7RATvpMOg6a+aeUCCUz0rFTSoDSKvST+ZRMoZRE+hl9a3yelFhJ15zSR5T6oO3pPOn+JykJczNpARZK32zdujXWqY0CLYST9k/nPsW1paQf3WNKt1BaKT239J5UFjUi9IzT+dD9oecqnSfdS3re6Bmi9y19ZnWRKvp/mJKE1UTnGH5TkCR1DgqSpM5BQZLUOShIkjoHBUlSNzp9RGkDSuukBT4obUBpCFpQ5r333hvUqMcPJRNSX5TWeCGT1GOkmj6qJKHo+CpJmNa4N0o6dkosUAKF0i20n3QvKPVB1/bPf/5zrD/77LOj9/3zn/881jdv3hzrlHpJzzOlieiZoERRSqZQwo6ecerlRJ+Z9kPPFe2b+hal86d9UOqQjruyn+r5UI8j+p915cqVQW1ubi5uS+8V/S9L53/27Nm4LT2Ha9asifV/5zcFSVLnoCBJ6hwUJEmdg4IkqXNQkCR1S76gWMSXfP/73491SomklddItV9MSvGMPI2v3PfatWtjPSUFqEcJJRzoPDds2DCoUWqI0l7VXifp2CkddujQoVifnZ2N9WpSLaFkBq0897Of/WxQo2vyk5/8JNYpTUXJlHSfqUcYJWco7ZauFSVeKPFD94Gu4V133TWo0bNcPZZUpzQeoXeCrm3anv5PUJ2SXXSeCT3LhBKG6XpRMo7u23PPPfeVn+83BUlS56AgSeocFCRJnYOCJKkbPdG8adOmWKfJLJpUraDJkin2QZPEpLp9RboFU31epf0FTajSpPwUx0iThzSRV5mEI/Rsrlq1avQ+brafpBqEWMznrXqfk//p96G12uRuFT2HdJ6V+0n7pkls+p+a3mXaBx1fasMx+Jyv3EKS9P+Gg4IkqXNQkCR1DgqSpM5BQZLUjY5PrF69OtbTYiCt1dIJi5lgouOgz6zM/FfbXJC0n+qiOZQ2oGNMSQ5K9tDCHPSZlcRXZcGX1lq7dOlSrKeFl+jc6VmmRXboGU9tNCghQ/etYookTGu8iM3MzMygVl1IilRaTlQTaVOkkugzqV5t0ZHQM077Tv8P6f1ZyPH5TUGS1DkoSJI6BwVJUuegIEnqHBQkSd3oeE91NjvNlFdTOZVEzVRJoGqiprJtJZWzmH1eqvunlASh80xJFkq33LhxI9apb086HzqOamKDFklJ20+RSlls1QVlKvuopnim2McU17zab6mSkKoeH31muj9TpNq+zG8KkqTOQUGS1DkoSJI6BwVJUuegIEnqFtxcqJIIoF5GlByifadUyVQrW1VWYKrug1T6KpEpetFU+yoRSrGkVEXqH9RaaytXroz1O++8c/RxUIqjuvJaJQ0yxWqBrdXuf7XvVcVipoyqz1VV+h+0mL2cWsvPSvUaVranbRdybf2mIEnqHBQkSZ2DgiSpc1CQJHUOCpKkbnT6qJqoSckH6gFSTWwsRr+Pf6F0QqW3TnXVtIpK/6TWOAk0RfKjmkBJdTq+anIm7buaDpufny99ZkoxVe8xJaEqvanoM2nflcRKtW9PtYdQZR90LJV9V3sZTZEQmuI9obrpI0nSonJQkCR1DgqSpM5BQZLUjZ5opoVJKpOtNPFHC6dMsW+q0yQcqWxfnYCe4nyq0kQUTcLRBF91kqsyUUaL7Hz44YexXnHt2rVYpzYXldYahAIMi7mwDd03et8qrVymmOCk543q1UnstP1U17BynlO1CqmcjxPNkqRJOChIkjoHBUlS56AgSeocFCRJ3aK1uUimaF3QWk5y0L4p9UGpgsr2lEiqpgoq51Ntc1FJj1STDFO0yqi0KLjZZ6ZrS9eq2kaBVD6TzrPS4qX6Dk5xP+n5meK+TZUymmLBn8qCXtVjmWqRnco+qvfn3/lNQZLUOShIkjoHBUlS56AgSeocFCRJ3YLTR5WkACV7aN+VRVIoJVBd2Ke64E9yK/oqVdNH6Rp+8sknpc8klcQGpSSoP0/1Wansg86f7s+t6FlVQe8m9TGbwmL2IVrMXltT7Ju2n2rfyUJSRsRvCpKkzkFBktQ5KEiSOgcFSVLnoCBJ6kbHXqqro1X6qFT7/FT2MUVPoOqxTJFgmup8KumjahKIVNITUyVQKkkgQukjui5p/1Pc+9Zq/W+qzzLdz3T+i7ny2hRJspvtJ5mqj9etXh1tMfftNwVJUuegIEnqHBQkSZ2DgiSpc1CQJHWj00d33HFHrFeSQ4u5mhj1VaomGSo9kaqfWTnP6kpy1fNM6QRK36xcubK070o/FkpxUELmxo0bo7ena0LP8vLly2N9ZmYm1m+77bZYT6qppEqihvZN+5ifn4/1ZcuWDWrV9FGl91F15bUpegVVV9ebIh1XXRluihXjFrJSpt8UJEmdg4IkqXNQkCR1DgqSpG70RHP1Z9OVn3uTKRYsqU5iL2SC5l9oMrjSLqG6DzLFgjeLOSFI15smBKdo/0DnQ585RZuLqVo6JNUWNGQxF8JJ13Cq1hJTtLm4FZ85xbFM0Srjy/ymIEnqHBQkSZ2DgiSpc1CQJHUOCpKkbsHpo0qKp/qTeTJF6mMx2w4QShQllHiZKk1VSYNUz30xf6Y/xTNUTYNQKmnp0uHrU33eqmmyCtr3YiZqKkm1qZJNi5kEWszU5a1IdI7hNwVJUuegIEnqHBQkSZ2DgiSpc1CQJHWj00fVBEpl4ZgpPrOa4qimRNKxVxMylWNMyZabfSahxFM6z6nSR1OkLaoJlMW8P5VU0hR9r+gzp0qaLOb5kMr21WRTZT9TpY8qz+dU/ZYqFvKs+E1BktQ5KEiSOgcFSVLnoCBJ6hwUJEnd6PTRFKrpgUoyY7F7HFWSU1P0LaJ+O2SKlMgUPXFuprLaW1UlfVSt0zGmhBjdt2ryrvLcVu9P5XyqiawpUkZT9ASi+mKmjMit6G+1kBX9/KYgSeocFCRJnYOCJKlzUJAkdQteZIcmOtLEZ3WBFKqnfVcniuhYpmhRQZO+dCzU0iKZ4j7Q9rdioplUn4m07+pkW3WCM30mXe8pFqmaqv0DqZzPFG1lpmqfUvnMW7HITnUBsMp1WciEMvGbgiSpc1CQJHUOCpKkzkFBktQ5KEiSugWnjyrpiWrLiSkWIKl+ZgUlM6oqLS2qnzlFm4tbsegJ7buSzKgmSug+0Gem7avvyWKaIjkz1eJVi3nfpmhzUU0pVlqiVNOIUyym5CI7kqRJOChIkjoHBUlS56AgSeocFCRJ3ZIvppjqliT9n+A3BUlS56AgSeocFCRJnYOCJKlzUJAkdQ4KkqTOQUGS1DkoSJI6BwVJUvcfEZ+d7bWNQb0AAAAASUVORK5CYII=",
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "import os\n",
        "import numpy as np\n",
        "from PIL import Image\n",
        "import torch\n",
        "from torchvision import transforms\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "# Initialize transformation\n",
        "transform = transforms.Compose([\n",
        "    transforms.ToTensor(),  # Convert PIL image to PyTorch tensor\n",
        "])\n",
        "\n",
        "# Folder path\n",
        "folder_path = 'images'\n",
        "\n",
        "# Read image file names from folder\n",
        "image_files = [f for f in os.listdir(folder_path) if f.endswith('.jpg')]\n",
        "\n",
        "# Initialize tensor for storing images\n",
        "tensor_list = []\n",
        "\n",
        "# Loop through image files to open, transform, and convert to tensor\n",
        "for image_file in image_files:\n",
        "    img_path = os.path.join(folder_path, image_file)\n",
        "    with Image.open(img_path) as img:\n",
        "        img_tensor = transform(img)\n",
        "        tensor_list.append(img_tensor)\n",
        "\n",
        "# Convert list of tensors to a single 4D tensor (Batch x Channels x Height x Width)\n",
        "image_tensor = torch.stack(tensor_list)\n",
        "\n",
        "# Print shape to confirm\n",
        "print(f\"Shape of the 4D tensor: {image_tensor.shape}\")\n",
        "\n",
        "# Show the last image using Matplotlib\n",
        "# First, remove the batch dimension and convert to numpy array\n",
        "last_image = image_tensor[-1].permute(1, 2, 0).numpy()\n",
        "\n",
        "# If the image is grayscale, remove the last dimension\n",
        "if last_image.shape[2] == 1:\n",
        "    last_image = np.squeeze(last_image, axis=2)\n",
        "\n",
        "plt.imshow(last_image, cmap='gray' if last_image.ndim == 2 else None)\n",
        "plt.axis('off')\n",
        "plt.show()\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mWa49PCNkhJI"
      },
      "source": [
        "*\"Let build a simple autoencoder and plot the resutls\"*\n",
        "\n",
        "output:"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 417
        },
        "id": "Jy2Is80oka9i",
        "outputId": "45a11af8-c99c-4f06-b5bb-86b0900fbcd4"
      },
      "outputs": [
        {
          "ename": "RuntimeError",
          "evalue": "ignored",
          "output_type": "error",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mRuntimeError\u001b[0m                              Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-3-c3a1f57d3b41>\u001b[0m in \u001b[0;36m<cell line: 57>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     58\u001b[0m     \u001b[0;32mfor\u001b[0m \u001b[0mbatch\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mdataloader\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     59\u001b[0m         \u001b[0mimgs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mbatch\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 60\u001b[0;31m         \u001b[0moutputs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mimgs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     61\u001b[0m         \u001b[0mloss\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mcriterion\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0moutputs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mimgs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     62\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py\u001b[0m in \u001b[0;36m_call_impl\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m   1499\u001b[0m                 \u001b[0;32mor\u001b[0m \u001b[0m_global_backward_pre_hooks\u001b[0m \u001b[0;32mor\u001b[0m \u001b[0m_global_backward_hooks\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1500\u001b[0m                 or _global_forward_hooks or _global_forward_pre_hooks):\n\u001b[0;32m-> 1501\u001b[0;31m             \u001b[0;32mreturn\u001b[0m \u001b[0mforward_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1502\u001b[0m         \u001b[0;31m# Do not call functions when jit is used\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1503\u001b[0m         \u001b[0mfull_backward_hooks\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnon_full_backward_hooks\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m<ipython-input-3-c3a1f57d3b41>\u001b[0m in \u001b[0;36mforward\u001b[0;34m(self, x)\u001b[0m\n\u001b[1;32m     43\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     44\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mforward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 45\u001b[0;31m         \u001b[0mx\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mencoder\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     46\u001b[0m         \u001b[0mx\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdecoder\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     47\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0mx\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py\u001b[0m in \u001b[0;36m_call_impl\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m   1499\u001b[0m                 \u001b[0;32mor\u001b[0m \u001b[0m_global_backward_pre_hooks\u001b[0m \u001b[0;32mor\u001b[0m \u001b[0m_global_backward_hooks\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1500\u001b[0m                 or _global_forward_hooks or _global_forward_pre_hooks):\n\u001b[0;32m-> 1501\u001b[0;31m             \u001b[0;32mreturn\u001b[0m \u001b[0mforward_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1502\u001b[0m         \u001b[0;31m# Do not call functions when jit is used\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1503\u001b[0m         \u001b[0mfull_backward_hooks\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnon_full_backward_hooks\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/torch/nn/modules/container.py\u001b[0m in \u001b[0;36mforward\u001b[0;34m(self, input)\u001b[0m\n\u001b[1;32m    215\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mforward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0minput\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    216\u001b[0m         \u001b[0;32mfor\u001b[0m \u001b[0mmodule\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 217\u001b[0;31m             \u001b[0minput\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmodule\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    218\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0minput\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    219\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/torch/nn/modules/module.py\u001b[0m in \u001b[0;36m_call_impl\u001b[0;34m(self, *args, **kwargs)\u001b[0m\n\u001b[1;32m   1499\u001b[0m                 \u001b[0;32mor\u001b[0m \u001b[0m_global_backward_pre_hooks\u001b[0m \u001b[0;32mor\u001b[0m \u001b[0m_global_backward_hooks\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1500\u001b[0m                 or _global_forward_hooks or _global_forward_pre_hooks):\n\u001b[0;32m-> 1501\u001b[0;31m             \u001b[0;32mreturn\u001b[0m \u001b[0mforward_call\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1502\u001b[0m         \u001b[0;31m# Do not call functions when jit is used\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1503\u001b[0m         \u001b[0mfull_backward_hooks\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mnon_full_backward_hooks\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m[\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/torch/nn/modules/conv.py\u001b[0m in \u001b[0;36mforward\u001b[0;34m(self, input)\u001b[0m\n\u001b[1;32m    461\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    462\u001b[0m     \u001b[0;32mdef\u001b[0m \u001b[0mforward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mself\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0minput\u001b[0m\u001b[0;34m:\u001b[0m \u001b[0mTensor\u001b[0m\u001b[0;34m)\u001b[0m \u001b[0;34m->\u001b[0m \u001b[0mTensor\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 463\u001b[0;31m         \u001b[0;32mreturn\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_conv_forward\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minput\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mweight\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mbias\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    464\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    465\u001b[0m \u001b[0;32mclass\u001b[0m \u001b[0mConv3d\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0m_ConvNd\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/torch/nn/modules/conv.py\u001b[0m in \u001b[0;36m_conv_forward\u001b[0;34m(self, input, weight, bias)\u001b[0m\n\u001b[1;32m    457\u001b[0m                             \u001b[0mweight\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mbias\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mstride\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    458\u001b[0m                             _pair(0), self.dilation, self.groups)\n\u001b[0;32m--> 459\u001b[0;31m         return F.conv2d(input, weight, bias, self.stride,\n\u001b[0m\u001b[1;32m    460\u001b[0m                         self.padding, self.dilation, self.groups)\n\u001b[1;32m    461\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mRuntimeError\u001b[0m: Given groups=1, weight of size [64, 3, 3, 3], expected input[8, 1, 64, 64] to have 3 channels, but got 1 channels instead"
          ]
        }
      ],
      "source": [
        "import os\n",
        "import torch\n",
        "import torch.nn as nn\n",
        "import torch.optim as optim\n",
        "from torchvision import transforms\n",
        "from torch.utils.data import DataLoader, TensorDataset\n",
        "import matplotlib.pyplot as plt\n",
        "import numpy as np\n",
        "from PIL import Image\n",
        "\n",
        "# Load Images (repeat code from before)\n",
        "transform = transforms.Compose([transforms.ToTensor()])\n",
        "folder_path = 'images'\n",
        "image_files = [f for f in os.listdir(folder_path) if f.endswith('.jpg')]\n",
        "tensor_list = []\n",
        "\n",
        "for image_file in image_files:\n",
        "    img_path = os.path.join(folder_path, image_file)\n",
        "    with Image.open(img_path) as img:\n",
        "        img_tensor = transform(img)\n",
        "        tensor_list.append(img_tensor)\n",
        "\n",
        "image_tensor = torch.stack(tensor_list)\n",
        "\n",
        "# Create DataLoader\n",
        "batch_size = 8\n",
        "dataset = TensorDataset(image_tensor)\n",
        "dataloader = DataLoader(dataset, batch_size=batch_size, shuffle=True)\n",
        "\n",
        "# Simple Autoencoder\n",
        "class Autoencoder(nn.Module):\n",
        "    def __init__(self):\n",
        "        super(Autoencoder, self).__init__()\n",
        "        self.encoder = nn.Sequential(\n",
        "            nn.Conv2d(3, 64, kernel_size=3, stride=1, padding=1),\n",
        "            nn.ReLU(),\n",
        "            nn.MaxPool2d(kernel_size=2, stride=2)\n",
        "        )\n",
        "        self.decoder = nn.Sequential(\n",
        "            nn.ConvTranspose2d(64, 3, kernel_size=2, stride=2),\n",
        "            nn.Sigmoid()\n",
        "        )\n",
        "\n",
        "    def forward(self, x):\n",
        "        x = self.encoder(x)\n",
        "        x = self.decoder(x)\n",
        "        return x\n",
        "\n",
        "# Initialize model, loss, and optimizer\n",
        "model = Autoencoder()\n",
        "criterion = nn.MSELoss()\n",
        "optimizer = optim.Adam(model.parameters(), lr=0.001)\n",
        "\n",
        "# Training\n",
        "n_epochs = 10\n",
        "\n",
        "for epoch in range(1, n_epochs + 1):\n",
        "    for batch in dataloader:\n",
        "        imgs, = batch\n",
        "        outputs = model(imgs)\n",
        "        loss = criterion(outputs, imgs)\n",
        "\n",
        "        optimizer.zero_grad()\n",
        "        loss.backward()\n",
        "        optimizer.step()\n",
        "\n",
        "    print(f\"Epoch [{epoch}/{n_epochs}], Loss: {loss.item():.4f}\")\n",
        "\n",
        "# Show the original and reconstructed images\n",
        "with torch.no_grad():\n",
        "    model.eval()\n",
        "    reconstructed = model(image_tensor[-1].unsqueeze(0))\n",
        "    reconstructed = reconstructed.squeeze(0).permute(1, 2, 0).numpy()\n",
        "\n",
        "original = image_tensor[-1].permute(1, 2, 0).numpy()\n",
        "\n",
        "fig, axes = plt.subplots(1, 2)\n",
        "axes[0].imshow(original)\n",
        "axes[0].set_title(\"Original\")\n",
        "axes[0].axis('off')\n",
        "\n",
        "axes[1].imshow(reconstructed)\n",
        "axes[1].set_title(\"Reconstructed\")\n",
        "axes[1].axis('off')\n",
        "\n",
        "plt.show()\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5CoPTjVolFLi"
      },
      "source": [
        "Turns out the images are grayscale and only have a single channel, so I just copy\n",
        "\"\n",
        "RuntimeError: Given groups=1, weight of size [64, 3, 3, 3], expected input[8, 1, 64, 64] to have 3 channels, but got 1 channels instead\" into chatGPT\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZLMlAhDwlYAj"
      },
      "source": [
        "Option 1: Grayscale Images\n",
        "Change the in_channels of nn.Conv2d to 1:\n",
        "\n",
        "python\n",
        "Copy code\n",
        "self.encoder = nn.Sequential(\n",
        "    nn.Conv2d(1, 64, kernel_size=3, stride=1, padding=1),  # Change here\n",
        "    nn.ReLU(),\n",
        "    nn.MaxPool2d(kernel_size=2, stride=2)\n",
        ")\n",
        "\n",
        "\n",
        "lets try that"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 518
        },
        "id": "aEPKmsPdlS21",
        "outputId": "535d1d1f-69e1-403a-ac43-4781fe0001fa"
      },
      "outputs": [
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/torch/nn/modules/loss.py:536: UserWarning: Using a target size (torch.Size([8, 1, 64, 64])) that is different to the input size (torch.Size([8, 3, 64, 64])). This will likely lead to incorrect results due to broadcasting. Please ensure they have the same size.\n",
            "  return F.mse_loss(input, target, reduction=self.reduction)\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch [1/10], Loss: 0.0571\n",
            "Epoch [2/10], Loss: 0.0191\n",
            "Epoch [3/10], Loss: 0.0073\n",
            "Epoch [4/10], Loss: 0.0056\n",
            "Epoch [5/10], Loss: 0.0028\n",
            "Epoch [6/10], Loss: 0.0035\n",
            "Epoch [7/10], Loss: 0.0022\n",
            "Epoch [8/10], Loss: 0.0022\n",
            "Epoch [9/10], Loss: 0.0032\n",
            "Epoch [10/10], Loss: 0.0031\n"
          ]
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 640x480 with 2 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "import os\n",
        "import torch\n",
        "import torch.nn as nn\n",
        "import torch.optim as optim\n",
        "from torchvision import transforms\n",
        "from torch.utils.data import DataLoader, TensorDataset\n",
        "import matplotlib.pyplot as plt\n",
        "import numpy as np\n",
        "from PIL import Image\n",
        "\n",
        "# Load Images (repeat code from before)\n",
        "transform = transforms.Compose([transforms.ToTensor()])\n",
        "folder_path = 'images'\n",
        "image_files = [f for f in os.listdir(folder_path) if f.endswith('.jpg')]\n",
        "tensor_list = []\n",
        "\n",
        "for image_file in image_files:\n",
        "    img_path = os.path.join(folder_path, image_file)\n",
        "    with Image.open(img_path) as img:\n",
        "        img_tensor = transform(img)\n",
        "        tensor_list.append(img_tensor)\n",
        "\n",
        "image_tensor = torch.stack(tensor_list)\n",
        "\n",
        "# Create DataLoader\n",
        "batch_size = 8\n",
        "dataset = TensorDataset(image_tensor)\n",
        "dataloader = DataLoader(dataset, batch_size=batch_size, shuffle=True)\n",
        "\n",
        "# Simple Autoencoder\n",
        "class Autoencoder(nn.Module):\n",
        "    def __init__(self):\n",
        "        super(Autoencoder, self).__init__()\n",
        "        self.encoder = nn.Sequential(\n",
        "            nn.Conv2d(1, 64, kernel_size=3, stride=1, padding=1),\n",
        "            nn.ReLU(),\n",
        "            nn.MaxPool2d(kernel_size=2, stride=2)\n",
        "        )\n",
        "        self.decoder = nn.Sequential(\n",
        "            nn.ConvTranspose2d(64, 3, kernel_size=2, stride=2),\n",
        "            nn.Sigmoid()\n",
        "        )\n",
        "\n",
        "    def forward(self, x):\n",
        "        x = self.encoder(x)\n",
        "        x = self.decoder(x)\n",
        "        return x\n",
        "\n",
        "# Initialize model, loss, and optimizer\n",
        "model = Autoencoder()\n",
        "criterion = nn.MSELoss()\n",
        "optimizer = optim.Adam(model.parameters(), lr=0.001)\n",
        "\n",
        "# Training\n",
        "n_epochs = 10\n",
        "\n",
        "for epoch in range(1, n_epochs + 1):\n",
        "    for batch in dataloader:\n",
        "        imgs, = batch\n",
        "        outputs = model(imgs)\n",
        "        loss = criterion(outputs, imgs)\n",
        "\n",
        "        optimizer.zero_grad()\n",
        "        loss.backward()\n",
        "        optimizer.step()\n",
        "\n",
        "    print(f\"Epoch [{epoch}/{n_epochs}], Loss: {loss.item():.4f}\")\n",
        "\n",
        "# Show the original and reconstructed images\n",
        "with torch.no_grad():\n",
        "    model.eval()\n",
        "    reconstructed = model(image_tensor[-1].unsqueeze(0))\n",
        "    reconstructed = reconstructed.squeeze(0).permute(1, 2, 0).numpy()\n",
        "\n",
        "original = image_tensor[-1].permute(1, 2, 0).numpy()\n",
        "\n",
        "fig, axes = plt.subplots(1, 2)\n",
        "axes[0].imshow(original)\n",
        "axes[0].set_title(\"Original\")\n",
        "axes[0].axis('off')\n",
        "\n",
        "axes[1].imshow(reconstructed)\n",
        "axes[1].set_title(\"Reconstructed\")\n",
        "axes[1].axis('off')\n",
        "\n",
        "plt.show()\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sXebhBmRlrcp"
      },
      "source": [
        "Works VERY well IF you know what you are doing."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "c55V186NlkMX"
      },
      "outputs": [],
      "source": []
    }
  ],
  "metadata": {
    "colab": {
      "authorship_tag": "ABX9TyPuv+MxHreKwl1gX4wYxMBp",
      "include_colab_link": true,
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "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": "3.9.7"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}