FashionMNIST_autoencoder.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "authorship_tag": "ABX9TyPYfnBBMHAx9CeTfVW2amZt",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/00ber/ml-projects/blob/main/FashionMNIST_autoencoder.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 12,
      "metadata": {
        "id": "NOhyoKFz3SP1"
      },
      "outputs": [],
      "source": [
        "import matplotlib.pyplot as plt\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import tensorflow as tf\n",
        "\n",
        "from sklearn.metrics import accuracy_score, precision_score, recall_score\n",
        "from sklearn.model_selection import train_test_split\n",
        "from tensorflow.keras import layers, losses\n",
        "from tensorflow.keras.datasets import fashion_mnist\n",
        "from tensorflow.keras.models import Model"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "(x_train, _), (x_test, _) = fashion_mnist.load_data()\n",
        "\n",
        "x_train = x_train.astype('float32') / 255.\n",
        "x_test = x_test.astype('float32') / 255.\n",
        "\n",
        "print (x_train.shape)\n",
        "print (x_test.shape)\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "B6S4K4cr3jQn",
        "outputId": "43fc0052-ace0-4c5d-b8a5-f70d2216f4ed"
      },
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "(60000, 28, 28)\n",
            "(10000, 28, 28)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "latent_dim_size = 60 \n",
        "\n",
        "class Autoencoder(Model):\n",
        "  def __init__(self, latent_dim_size):\n",
        "    super(Autoencoder, self).__init__()\n",
        "    self.latent_dim_size = latent_dim_size   \n",
        "    self.encoder = tf.keras.Sequential([\n",
        "        tf.keras.layers.InputLayer(input_shape=(28, 28, 1)),\n",
        "        tf.keras.layers.Conv2D(32, 1, activation='relu'),\n",
        "        tf.keras.layers.Flatten(),\n",
        "        tf.keras.layers.Dense(latent_dim_size)\n",
        "    ])\n",
        "    self.decoder = tf.keras.Sequential([\n",
        "        layers.InputLayer(input_shape=(latent_dim_size,)),\n",
        "        layers.Dense(784, activation='relu'),\n",
        "        layers.Reshape((28, 28))\n",
        "    ])\n",
        "\n",
        "  def call(self, x):\n",
        "    encoded = self.encoder(x)\n",
        "    decoded = self.decoder(encoded)\n",
        "    return decoded\n",
        "\n",
        "autoencoder = Autoencoder(latent_dim_size)"
      ],
      "metadata": {
        "id": "2IX5KqnC3w2C"
      },
      "execution_count": 43,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "autoencoder.compile(optimizer='adam', loss=losses.MeanSquaredError())"
      ],
      "metadata": {
        "id": "aeGRChGt4a_4"
      },
      "execution_count": 44,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "autoencoder.fit(x_train, x_train,\n",
        "                epochs=20,\n",
        "                shuffle=True,\n",
        "                validation_data=(x_test, x_test))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "BvdsHnxb7n4H",
        "outputId": "bdf44a2b-9d17-41a3-a0a5-12da74061092"
      },
      "execution_count": 45,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/20\n",
            "1875/1875 [==============================] - 71s 37ms/step - loss: 0.0193 - val_loss: 0.0117\n",
            "Epoch 2/20\n",
            "1875/1875 [==============================] - 66s 35ms/step - loss: 0.0109 - val_loss: 0.0104\n",
            "Epoch 3/20\n",
            "1875/1875 [==============================] - 66s 35ms/step - loss: 0.0099 - val_loss: 0.0096\n",
            "Epoch 4/20\n",
            "1875/1875 [==============================] - 66s 35ms/step - loss: 0.0095 - val_loss: 0.0094\n",
            "Epoch 5/20\n",
            "1875/1875 [==============================] - 68s 36ms/step - loss: 0.0094 - val_loss: 0.0094\n",
            "Epoch 6/20\n",
            "1875/1875 [==============================] - 69s 37ms/step - loss: 0.0093 - val_loss: 0.0093\n",
            "Epoch 7/20\n",
            "1875/1875 [==============================] - 67s 36ms/step - loss: 0.0093 - val_loss: 0.0093\n",
            "Epoch 8/20\n",
            "1875/1875 [==============================] - 68s 36ms/step - loss: 0.0092 - val_loss: 0.0093\n",
            "Epoch 9/20\n",
            "1875/1875 [==============================] - 69s 37ms/step - loss: 0.0090 - val_loss: 0.0089\n",
            "Epoch 10/20\n",
            "1875/1875 [==============================] - 68s 36ms/step - loss: 0.0089 - val_loss: 0.0090\n",
            "Epoch 11/20\n",
            "1875/1875 [==============================] - 67s 36ms/step - loss: 0.0088 - val_loss: 0.0089\n",
            "Epoch 12/20\n",
            "1875/1875 [==============================] - 67s 36ms/step - loss: 0.0088 - val_loss: 0.0089\n",
            "Epoch 13/20\n",
            "1875/1875 [==============================] - 63s 34ms/step - loss: 0.0088 - val_loss: 0.0088\n",
            "Epoch 14/20\n",
            "1875/1875 [==============================] - 68s 36ms/step - loss: 0.0088 - val_loss: 0.0088\n",
            "Epoch 15/20\n",
            "1875/1875 [==============================] - 67s 36ms/step - loss: 0.0087 - val_loss: 0.0088\n",
            "Epoch 16/20\n",
            "1875/1875 [==============================] - 72s 38ms/step - loss: 0.0087 - val_loss: 0.0088\n",
            "Epoch 17/20\n",
            "1875/1875 [==============================] - 69s 37ms/step - loss: 0.0087 - val_loss: 0.0088\n",
            "Epoch 18/20\n",
            "1875/1875 [==============================] - 67s 36ms/step - loss: 0.0087 - val_loss: 0.0088\n",
            "Epoch 19/20\n",
            "1875/1875 [==============================] - 65s 35ms/step - loss: 0.0087 - val_loss: 0.0088\n",
            "Epoch 20/20\n",
            "1875/1875 [==============================] - 66s 35ms/step - loss: 0.0087 - val_loss: 0.0088\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x7f9edc4a5340>"
            ]
          },
          "metadata": {},
          "execution_count": 45
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "encoded_imgs = autoencoder.encoder(x_test).numpy()\n",
        "decoded_imgs = autoencoder.decoder(encoded_imgs).numpy()"
      ],
      "metadata": {
        "id": "FvhbvNfP7s5M"
      },
      "execution_count": 46,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "n = 10\n",
        "plt.figure(figsize=(20, 4))\n",
        "for i in range(n):\n",
        "  # display original\n",
        "  ax = plt.subplot(2, n, i + 1)\n",
        "  plt.imshow(x_test[i])\n",
        "  plt.title(\"original\")\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(2, n, i + 1 + n)\n",
        "  plt.imshow(decoded_imgs[i])\n",
        "  plt.title(\"reconstructed\")\n",
        "  plt.gray()\n",
        "  ax.get_xaxis().set_visible(False)\n",
        "  ax.get_yaxis().set_visible(False)\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 264
        },
        "id": "cUwUb34U_npk",
        "outputId": "37430b51-e586-4055-b6a6-e519ead65c0f"
      },
      "execution_count": 47,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1440x288 with 20 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "X5Giynzn_p-B"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}