06_autoencoder.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "colab": {
      "name": "Copy of 06_autoencoder.ipynb",
      "provenance": []
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZvJfBaRn3P42",
        "colab_type": "text"
      },
      "source": [
        "# Autoencoders\n",
        "\n",
        "[Source](https://twitter.com/rickwierenga/status/1216801014004797446)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZDurKWtZd2ZK",
        "colab_type": "code",
        "outputId": "f7b7443d-77b4-4517-d2f9-b750d9507e48",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "%tensorflow_version 2.x"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "TensorFlow 2.x selected.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "se3XrQkXfECV",
        "colab_type": "code",
        "outputId": "fa3666b7-e83e-4d17-f447-9a9a95dd5f05",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "%pylab inline"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Populating the interactive namespace from numpy and matplotlib\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mokbCApx3IL0",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import tensorflow as tf"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UA4GoZEG4IB0",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "(x_train, _), (x_test, _) = tf.keras.datasets.mnist.load_data()\n",
        "x_train = x_train / 255\n",
        "x_test = x_test / 255"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nzMVAw8R3nHE",
        "colab_type": "text"
      },
      "source": [
        "# Simple auto encoder"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2kWUdt_frO5q",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "encoder = tf.keras.models.Sequential([\n",
        "    tf.keras.layers.Flatten(input_shape=[28, 28]),\n",
        "    tf.keras.layers.Dense(100, activation=\"relu\"),\n",
        "    tf.keras.layers.Dense(30, activation=\"relu\"),\n",
        "])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MlY5WNZbrSlS",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "decoder = tf.keras.models.Sequential([\n",
        "    tf.keras.layers.Dense(100, activation=\"relu\", input_shape=[30]),\n",
        "    tf.keras.layers.Dense(28 * 28, activation=\"sigmoid\"),\n",
        "    tf.keras.layers.Reshape([28, 28])\n",
        "])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wvcHl-xx4Az2",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "stacked_autoencoder = tf.keras.models.Sequential([encoder, decoder])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vvREkPRD4DvD",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "stacked_autoencoder.compile(loss=\"binary_crossentropy\",\n",
        "                   optimizer='adam')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IqaTTymB4Fq_",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 374
        },
        "outputId": "78654578-1643-41b2-b60c-2d343f930ecc"
      },
      "source": [
        "history = stacked_autoencoder.fit(x_train, x_train, epochs=10,\n",
        "                                  validation_data=[x_test, x_test])"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Train on 60000 samples, validate on 10000 samples\n",
            "Epoch 1/10\n",
            "60000/60000 [==============================] - 6s 94us/sample - loss: 0.1448 - val_loss: 0.1079\n",
            "Epoch 2/10\n",
            "60000/60000 [==============================] - 5s 86us/sample - loss: 0.1042 - val_loss: 0.0989\n",
            "Epoch 3/10\n",
            "60000/60000 [==============================] - 5s 83us/sample - loss: 0.0988 - val_loss: 0.0960\n",
            "Epoch 4/10\n",
            "60000/60000 [==============================] - 5s 83us/sample - loss: 0.0960 - val_loss: 0.0933\n",
            "Epoch 5/10\n",
            "60000/60000 [==============================] - 5s 84us/sample - loss: 0.0936 - val_loss: 0.0922\n",
            "Epoch 6/10\n",
            "60000/60000 [==============================] - 5s 82us/sample - loss: 0.0921 - val_loss: 0.0908\n",
            "Epoch 7/10\n",
            "60000/60000 [==============================] - 5s 82us/sample - loss: 0.0911 - val_loss: 0.0897\n",
            "Epoch 8/10\n",
            "60000/60000 [==============================] - 5s 83us/sample - loss: 0.0903 - val_loss: 0.0888\n",
            "Epoch 9/10\n",
            "60000/60000 [==============================] - 5s 82us/sample - loss: 0.0896 - val_loss: 0.0883\n",
            "Epoch 10/10\n",
            "60000/60000 [==============================] - 5s 83us/sample - loss: 0.0891 - val_loss: 0.0880\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "u-4mTsF94F16",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 317
        },
        "outputId": "55bd9ee2-2693-4d44-aebf-ad6eeef136b3"
      },
      "source": [
        "figsize(20, 5)\n",
        "for i in range(8):\n",
        "  subplot(2, 8, i+1)\n",
        "  pred = stacked_autoencoder.predict(x_test[i].reshape((1, 28, 28)))\n",
        "  imshow(x_test[i], cmap=\"binary\")\n",
        "\n",
        "  subplot(2, 8, i+8+1)\n",
        "  imshow(pred.reshape((28, 28)), cmap=\"binary\")"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAABH4AAAEsCAYAAAC1y3RCAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nWu09Aeq23cUAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 1440x360 with 16 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9uUNg-cT4QlF",
        "colab_type": "text"
      },
      "source": [
        "## Convolutional auto encoder"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rHUxc2AC4PMi",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "encoder = tf.keras.models.Sequential([\n",
        "    tf.keras.layers.Reshape([28, 28, 1], input_shape=[28, 28]),\n",
        "    tf.keras.layers.Conv2D(16, kernel_size=(3, 3), padding=\"same\", activation=\"relu\"),\n",
        "    tf.keras.layers.MaxPool2D(pool_size=2),\n",
        "    tf.keras.layers.Conv2D(32, kernel_size=(3, 3), padding=\"same\", activation=\"relu\"),\n",
        "    tf.keras.layers.MaxPool2D(pool_size=2),\n",
        "    tf.keras.layers.Conv2D(64, kernel_size=(3, 3), padding=\"same\", activation=\"relu\"),\n",
        "    tf.keras.layers.MaxPool2D(pool_size=2)\n",
        "])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jGg5m5Ey4VvF",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "decoder = tf.keras.models.Sequential([\n",
        "    tf.keras.layers.Conv2DTranspose(32, kernel_size=(3, 3), strides=2, padding=\"valid\",\n",
        "                                    activation=\"relu\",\n",
        "                                    input_shape=[3, 3, 64]),\n",
        "    tf.keras.layers.Conv2DTranspose(16, kernel_size=(3, 3), strides=2, padding=\"same\",\n",
        "                                    activation=\"relu\"),\n",
        "    tf.keras.layers.Conv2DTranspose(1, kernel_size=(3, 3), strides=2, padding=\"same\",\n",
        "                                    activation=\"sigmoid\"),\n",
        "    tf.keras.layers.Reshape([28, 28])\n",
        "])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VbexhKjN1vBX",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "stacked_autoencoder = tf.keras.models.Sequential([encoder, decoder])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "30DUbxUk1r9S",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "stacked_autoencoder.compile(loss=\"binary_crossentropy\",\n",
        "                            optimizer='adam')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NCstnS7BeNis",
        "colab_type": "code",
        "outputId": "8a8414ae-92d5-493a-d7a0-0cc21c1c6ffa",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 374
        }
      },
      "source": [
        "history = stacked_autoencoder.fit(x_train, x_train, epochs=10,\n",
        "                         validation_data=[x_test, x_test])"
      ],
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Train on 60000 samples, validate on 10000 samples\n",
            "Epoch 1/10\n",
            "60000/60000 [==============================] - 9s 152us/sample - loss: 0.1179 - val_loss: 0.0835\n",
            "Epoch 2/10\n",
            "60000/60000 [==============================] - 7s 112us/sample - loss: 0.0805 - val_loss: 0.0774\n",
            "Epoch 3/10\n",
            "60000/60000 [==============================] - 7s 112us/sample - loss: 0.0764 - val_loss: 0.0742\n",
            "Epoch 4/10\n",
            "60000/60000 [==============================] - 7s 111us/sample - loss: 0.0742 - val_loss: 0.0724\n",
            "Epoch 5/10\n",
            "60000/60000 [==============================] - 7s 113us/sample - loss: 0.0728 - val_loss: 0.0717\n",
            "Epoch 6/10\n",
            "60000/60000 [==============================] - 7s 110us/sample - loss: 0.0718 - val_loss: 0.0708\n",
            "Epoch 7/10\n",
            "60000/60000 [==============================] - 7s 111us/sample - loss: 0.0711 - val_loss: 0.0704\n",
            "Epoch 8/10\n",
            "60000/60000 [==============================] - 7s 111us/sample - loss: 0.0706 - val_loss: 0.0700\n",
            "Epoch 9/10\n",
            "60000/60000 [==============================] - 7s 111us/sample - loss: 0.0701 - val_loss: 0.0693\n",
            "Epoch 10/10\n",
            "60000/60000 [==============================] - 7s 111us/sample - loss: 0.0697 - val_loss: 0.0689\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KhdR1YjS4dur",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 317
        },
        "outputId": "12ba74c8-9a9a-4487-8730-a715b7b74063"
      },
      "source": [
        "figsize(20, 5)\n",
        "for i in range(8):\n",
        "  subplot(2, 8, i+1)\n",
        "  pred = stacked_autoencoder.predict(x_test[i].reshape((1, 28, 28)))\n",
        "  imshow(x_test[i], cmap=\"binary\")\n",
        "\n",
        "  subplot(2, 8, i+8+1)\n",
        "  imshow(pred.reshape((28, 28)), cmap=\"binary\")"
      ],
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAABH4AAAEsCAYAAAC1y3RCAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAAYur/A34+2apb3EkgAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 1440x360 with 16 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0hSe9lFR4k7z",
        "colab_type": "text"
      },
      "source": [
        "## Denoising auto encoder"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZWj3gHZ04mJT",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "encoder = tf.keras.models.Sequential([\n",
        "    tf.keras.layers.Flatten(input_shape=[28, 28]),\n",
        "    tf.keras.layers.Dense(100, activation=\"relu\"),\n",
        "    tf.keras.layers.Dense(100, activation=\"relu\"),\n",
        "    tf.keras.layers.Dense(30, activation=\"relu\")\n",
        "])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OibuvIQc4ho5",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "decoder = tf.keras.models.Sequential([\n",
        "    tf.keras.layers.Dense(100, activation=\"relu\", input_shape=[30]),\n",
        "    tf.keras.layers.Dense(100, activation=\"relu\"),\n",
        "    tf.keras.layers.Dense(28 * 28, activation=\"sigmoid\"),\n",
        "    tf.keras.layers.Reshape([28, 28])\n",
        "])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mZTPubqp4qte",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "stacked_autoencoder = tf.keras.models.Sequential([encoder, decoder])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HFvxnafB4r0b",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "stacked_autoencoder.compile(loss=\"binary_crossentropy\",\n",
        "                            optimizer='adam')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PZ1T_ofW4sus",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "x_train_noise = x_train + ((np.random.random(x_train.shape)) / 4)\n",
        "x_test_noise = x_test + ((np.random.random(x_test.shape)) / 4)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NyPLjcYr4uED",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 374
        },
        "outputId": "d0a5d4ca-a9f3-42a8-c565-31455912e9eb"
      },
      "source": [
        "history = stacked_autoencoder.fit(x_train_noise, x_train, epochs=10,\n",
        "                                  validation_data=[x_test_noise, x_test])"
      ],
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Train on 60000 samples, validate on 10000 samples\n",
            "Epoch 1/10\n",
            "60000/60000 [==============================] - 6s 97us/sample - loss: 0.1631 - val_loss: 0.1248\n",
            "Epoch 2/10\n",
            "60000/60000 [==============================] - 5s 90us/sample - loss: 0.1194 - val_loss: 0.1121\n",
            "Epoch 3/10\n",
            "60000/60000 [==============================] - 5s 90us/sample - loss: 0.1100 - val_loss: 0.1052\n",
            "Epoch 4/10\n",
            "60000/60000 [==============================] - 6s 93us/sample - loss: 0.1051 - val_loss: 0.1020\n",
            "Epoch 5/10\n",
            "60000/60000 [==============================] - 5s 89us/sample - loss: 0.1018 - val_loss: 0.0995\n",
            "Epoch 6/10\n",
            "60000/60000 [==============================] - 5s 89us/sample - loss: 0.0996 - val_loss: 0.0971\n",
            "Epoch 7/10\n",
            "60000/60000 [==============================] - 5s 89us/sample - loss: 0.0982 - val_loss: 0.0964\n",
            "Epoch 8/10\n",
            "60000/60000 [==============================] - 5s 89us/sample - loss: 0.0973 - val_loss: 0.0955\n",
            "Epoch 9/10\n",
            "60000/60000 [==============================] - 5s 89us/sample - loss: 0.0965 - val_loss: 0.0967\n",
            "Epoch 10/10\n",
            "60000/60000 [==============================] - 5s 89us/sample - loss: 0.0959 - val_loss: 0.0946\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "853h_kqq4voD",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 317
        },
        "outputId": "46e26231-13c7-4679-ae99-c7c6697da6c6"
      },
      "source": [
        "figsize(20, 5)\n",
        "for i in range(8):\n",
        "  subplot(2, 8, i+1)\n",
        "  imshow(x_test_noise[i], cmap=\"binary\")\n",
        "\n",
        "  subplot(2, 8, i+8+1)\n",
        "  pred = stacked_autoencoder.predict(x_test_noise[i].reshape((1, 28, 28)))\n",
        "  imshow(pred.reshape((28, 28)), cmap=\"binary\")"
      ],
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAABH4AAAEsCAYAAAC1y3RCAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 1440x360 with 16 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    }
  ]
}