1-imdb-binary-classification/one_hot.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "one-hot.ipynb",
      "version": "0.3.2",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/unclepeddy/deeplearning/blob/master/1-imdb-binary-classification/one_hot.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "metadata": {
        "id": "XRXJZ-M-iO0H",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "outputId": "a2ebe634-93e7-4d81-a685-a76dfa57d99a"
      },
      "cell_type": "code",
      "source": [
        "! import tf-nightly-2.0-preview"
      ],
      "execution_count": 0,
      "outputs": [
      ]
    },
    {
      "metadata": {
        "id": "j63UeAUHiXwl",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "outputId": "3516f5b2-f048-4bd9-f624-9fda0a4e4fae"
      },
      "cell_type": "code",
      "source": [
        "import tensorflow as tf\n",
        "\n",
        "from tensorflow.keras import metrics, models, layers, losses, optimizers\n",
        "from tensorflow.keras.datasets import imdb\n",
        "\n",
        "import matplotlib.pyplot as plt\n",
        "import numpy as np\n",
        "\n",
        "print(tf.__version__)"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "2.0.0-dev20190330\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "metadata": {
        "id": "vkX8WPL-iV4B",
        "colab_type": "code",
        "colab": {}
      },
      "cell_type": "code",
      "source": [
        "# Restrict dataset to NUM_WORDS most frequent words\n",
        "num_words = 10000\n",
        "# Size of validation set\n",
        "validation_size = 10000\n",
        "\n",
        "# Import labeled traning and test data\n",
        "(train_data, train_labels), (test_data, test_labels) = imdb.load_data(\n",
        "    num_words=num_words)\n",
        "\n",
        "# Given list of word sequences, return list of one-hot encoded lists\n",
        "def vectorize_sequences(sequences, dimension=num_words):\n",
        "  results = np.zeros((len(sequences), dimension))\n",
        "  for i, sequence in enumerate(sequences):\n",
        "    results[i, sequence] = 1\n",
        "  return results\n",
        "\n",
        "# Vectorize the training and test data\n",
        "x_train = vectorize_sequences(train_data)\n",
        "x_test = vectorize_sequences(test_data)\n",
        "y_train = np.asarray(train_labels).astype('float32')\n",
        "y_test = np.asarray(test_labels).astype('float32')\n",
        "\n",
        "# Set aside validation set\n",
        "x_val = x_train[:validation_size]\n",
        "y_val = y_train[:validation_size]\n",
        "partial_x_train = x_train[validation_size:]\n",
        "partial_y_train = y_train[validation_size:]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "metadata": {
        "id": "7d6_eMt_iV5s",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 933
        },
        "outputId": "e9c9c463-1aae-41cd-b24c-c163a1fbc2be"
      },
      "cell_type": "code",
      "source": [
        "# Build simple model: dense w/ ReLU -> dense w/ ReLU -> sigmoid\n",
        "model = models.Sequential()\n",
        "model.add(layers.Dense(16, activation='relu', input_shape=(num_words,)))\n",
        "model.add(layers.Dense(16, activation='relu'))\n",
        "model.add(layers.Dense(1, activation='sigmoid'))\n",
        "\n",
        "# Compile and train the model\n",
        "model.compile(optimizer=optimizers.RMSprop(lr=0.001),\n",
        "             loss=losses.binary_crossentropy,\n",
        "             metrics=[metrics.binary_accuracy])\n",
        "history = model.fit(partial_x_train,\n",
        "                   partial_y_train,\n",
        "                   epochs=4,\n",
        "                   batch_size=512,\n",
        "                   validation_data=(x_val, y_val))\n",
        "\n",
        "# Capture training statistics\n",
        "history_dict = history.history\n",
        "loss_values = history_dict['loss']\n",
        "val_loss_values = history_dict['val_loss']\n",
        "binary_accuracy = history_dict['binary_accuracy']\n",
        "val_binary_accuracy = history_dict['val_binary_accuracy']\n",
        "\n",
        "epochs = range(1, len(history_dict.get('binary_accuracy')) + 1)\n",
        "\n",
        "# Plot the training and validation loss\n",
        "plt.plot(epochs, loss_values, 'bo', label='Training loss')\n",
        "plt.plot(epochs, val_loss_values, 'b', label='Validation loss')\n",
        "plt.title('Training and Validation Loss')\n",
        "plt.xlabel('Epochs')\n",
        "plt.ylabel('Loss')\n",
        "plt.legend()\n",
        "plt.show()\n",
        "\n",
        "# Plot the training and validation accuracy\n",
        "plt.plot(epochs, binary_accuracy, 'bo', label='Training accuracy')\n",
        "plt.plot(epochs, val_binary_accuracy, 'b', label='Validation accuracy')\n",
        "plt.title('Training and Validation Accuracy')\n",
        "plt.xlabel('Epochs')\n",
        "plt.ylabel('Loss')\n",
        "plt.legend()\n",
        "plt.show()\n",
        "\n",
        "# Evaluate the model with the test data\n",
        "loss, accuracy = model.evaluate(x_test, y_test)\n",
        "print(loss, accuracy)"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Train on 15000 samples, validate on 10000 samples\n",
            "Epoch 1/4\n",
            "15000/15000 [==============================] - 4s 281us/sample - loss: 0.5217 - binary_accuracy: 0.7838 - val_loss: 0.3912 - val_binary_accuracy: 0.8615\n",
            "Epoch 2/4\n",
            "15000/15000 [==============================] - 3s 210us/sample - loss: 0.3087 - binary_accuracy: 0.8985 - val_loss: 0.3078 - val_binary_accuracy: 0.8837\n",
            "Epoch 3/4\n",
            "15000/15000 [==============================] - 3s 216us/sample - loss: 0.2261 - binary_accuracy: 0.9255 - val_loss: 0.2782 - val_binary_accuracy: 0.8917\n",
            "Epoch 4/4\n",
            "15000/15000 [==============================] - 3s 213us/sample - loss: 0.1821 - binary_accuracy: 0.9404 - val_loss: 0.2783 - val_binary_accuracy: 0.8886\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfUAAAFnCAYAAAC/5tBZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 576x396 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfUAAAFnCAYAAAC/5tBZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nlyQpEoa6JEmRMNQlSYrEB5I6yeQYeUB9AAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 576x396 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "25000/25000 [==============================] - 4s 169us/sample - loss: 0.2976 - binary_accuracy: 0.8783\n",
            "0.2976436975097656 0.87832\n"
          ],
          "name": "stdout"
        }
      ]
    }
  ]
}