Class 3 - Autoencoder.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Autoencoder.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "PiIDDU_iVVsG",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "%tensorflow_version 2.x\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "import tensorflow as tf"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GF3ongiwCjIo",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "mnist = tf.keras.datasets.mnist # mnist data of grayscale images of handwritten digits\n",
        "(x_train, y_train), (x_test, y_test) = mnist.load_data() # load it into training and testing sets\n",
        "x_train, x_test = x_train/255.0, x_test/255.0 # the data is in the range (0,255) grayscale value for each pixel, normalize it"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rRd0_7ktI0Fd",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "print(x_train.shape, x_test.shape) # look at how many images are there"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MMa4n9DUI25i",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "print(y_train.shape, y_test.shape)\n",
        "print(y_train[:10])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rCrAKCRUS64a",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# plot the first ten images\n",
        "for i in range(10):\n",
        "    plt.figure()\n",
        "    plt.imshow(x_train[i], cmap='gray')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "u0zNDSlJguBM",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#---------------------------------------------------\n",
        "# build an autoencoder model\n",
        "def create_modelAE():\n",
        "    # build a sequential AutoEncoder model\n",
        "    model = tf.keras.models.Sequential([\n",
        "        # the input layer                               \n",
        "        tf.keras.layers.InputLayer((28,28)),\n",
        "        # flatten it into a single column\n",
        "        tf.keras.layers.Flatten(), \n",
        "        # bottleneck layer\n",
        "        tf.keras.layers.Dense(32, activation='relu'),\n",
        "        # the output layer\n",
        "        tf.keras.layers.Dense(28*28, activation='sigmoid'),\n",
        "        # the output layer, reshape the flattened column into an image shape\n",
        "\t      tf.keras.layers.Reshape((28, 28, 1))\n",
        "\t  ])\n",
        "    return model"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6yAtvAd6izhP",
        "colab_type": "code",
        "outputId": "291990ad-ed7a-4a30-abfa-bcd50028eab7",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 332
        }
      },
      "source": [
        "modelAE = create_modelAE()\n",
        "print(modelAE.summary())"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "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, 32)                25120     \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 784)               25872     \n",
            "_________________________________________________________________\n",
            "reshape (Reshape)            (None, 28, 28, 1)         0         \n",
            "=================================================================\n",
            "Total params: 50,992\n",
            "Trainable params: 50,992\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n",
            "None\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Ql_eSskiJ2y4",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "500*500*3*32 + 32"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Djti_AdHlHoj",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model = modelAE"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CqFlPJkxgU2H",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# compile, using the adamax optimizer for gradient descend and mean square error for loss\n",
        "model.compile(optimizer='adamax', loss='mse')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tWyYxt3klDq3",
        "colab_type": "code",
        "outputId": "4a218475-b2ed-4903-e3ab-c35232ac0146",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 384
        }
      },
      "source": [
        "# fit (train) the model on the test data. The input and the output is the same because\n",
        "# we are trying to re-construct the original input after compression\n",
        "train_hist = model.fit(x_train, x_train, epochs=10)                        "
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/10\n",
            "1875/1875 [==============================] - 3s 2ms/step - loss: 0.0463\n",
            "Epoch 2/10\n",
            "1875/1875 [==============================] - 3s 2ms/step - loss: 0.0224\n",
            "Epoch 3/10\n",
            "1875/1875 [==============================] - 3s 2ms/step - loss: 0.0159\n",
            "Epoch 4/10\n",
            "1875/1875 [==============================] - 3s 2ms/step - loss: 0.0130\n",
            "Epoch 5/10\n",
            "1875/1875 [==============================] - 3s 2ms/step - loss: 0.0117\n",
            "Epoch 6/10\n",
            "1875/1875 [==============================] - 4s 2ms/step - loss: 0.0111\n",
            "Epoch 7/10\n",
            "1875/1875 [==============================] - 4s 2ms/step - loss: 0.0108\n",
            "Epoch 8/10\n",
            "1875/1875 [==============================] - 4s 2ms/step - loss: 0.0106\n",
            "Epoch 9/10\n",
            "1875/1875 [==============================] - 4s 2ms/step - loss: 0.0105\n",
            "Epoch 10/10\n",
            "1875/1875 [==============================] - 4s 2ms/step - loss: 0.0104\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gdqVpHsTlgNh",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "#---------------------------------------------------\n",
        "# plots n images in two rows, the first row is the original images\n",
        "# and the second row is the reconstructed (decoded) images\n",
        "def show_results(n):\n",
        "    # plot\n",
        "    plt.figure(figsize=(12, 12))\n",
        "    for i in range(n):\n",
        "        # display original image\n",
        "        ax = plt.subplot(3, n, i + 1)\n",
        "        ax.set_title(\"Original\")\n",
        "        plt.imshow(original[i])\n",
        "        plt.gray()\n",
        "        ax.get_xaxis().set_visible(False)\n",
        "        ax.get_yaxis().set_visible(False)\n",
        "\n",
        "        # display reconstruction\n",
        "        ax = plt.subplot(3, n, i + 1 + n)\n",
        "        ax.set_title(\"Decoded\")\n",
        "        plt.imshow(decoded[i])\n",
        "        plt.gray()\n",
        "        ax.get_xaxis().set_visible(False)\n",
        "        ax.get_yaxis().set_visible(False)\n",
        "    plt.show()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KBDl49xjOHsJ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "original = []\n",
        "decoded = []\n",
        "n=0\n",
        "# predict the first 10 items from the test set\n",
        "for img in x_test[:10]:\n",
        "    #predict\n",
        "    pimg = model.predict(tf.expand_dims(img, 0))\n",
        "    original.append(img) # original\n",
        "    decoded.append(pimg.squeeze()) # predicted\n",
        "    n += 1"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8Azgx94TO-2q",
        "colab_type": "code",
        "outputId": "7e1c3ab8-3e39-4e31-d39f-3ae807d6c140",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 358
        }
      },
      "source": [
        "show_results(n)"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 864x864 with 20 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uT6E9d60jmho",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "28*28/32"
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}