TCN_IMDB.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "TCN-IMDB.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyMKeU+qGJIVBwfbnA27BbAJ",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/ashishpatel26/tcn-keras-Examples/blob/master/TCN_IMDB.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UKoPuUXYFlQ_",
        "colab_type": "text"
      },
      "source": [
        "### Load the Packages"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kvqRJevMGz18",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 170
        },
        "outputId": "c6889333-80d6-4c9c-8cdb-72ba6c93a54c"
      },
      "source": [
        "!pip install keras-tcn"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Requirement already satisfied: keras-tcn in /usr/local/lib/python3.6/dist-packages (3.1.1)\n",
            "Requirement already satisfied: keras==2.3.1 in /usr/local/lib/python3.6/dist-packages (from keras-tcn) (2.3.1)\n",
            "Requirement already satisfied: numpy>=1.18.1 in /usr/local/lib/python3.6/dist-packages (from keras-tcn) (1.18.5)\n",
            "Requirement already satisfied: scipy>=0.14 in /usr/local/lib/python3.6/dist-packages (from keras==2.3.1->keras-tcn) (1.4.1)\n",
            "Requirement already satisfied: h5py in /usr/local/lib/python3.6/dist-packages (from keras==2.3.1->keras-tcn) (2.10.0)\n",
            "Requirement already satisfied: pyyaml in /usr/local/lib/python3.6/dist-packages (from keras==2.3.1->keras-tcn) (3.13)\n",
            "Requirement already satisfied: keras-applications>=1.0.6 in /usr/local/lib/python3.6/dist-packages (from keras==2.3.1->keras-tcn) (1.0.8)\n",
            "Requirement already satisfied: keras-preprocessing>=1.0.5 in /usr/local/lib/python3.6/dist-packages (from keras==2.3.1->keras-tcn) (1.1.2)\n",
            "Requirement already satisfied: six>=1.9.0 in /usr/local/lib/python3.6/dist-packages (from keras==2.3.1->keras-tcn) (1.15.0)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kfPWht1ZCoSU",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import numpy as np\n",
        "from tensorflow.keras import Model, Input\n",
        "from tensorflow.keras.datasets import imdb\n",
        "from tensorflow.keras.layers import Dense, Dropout, Embedding\n",
        "from tensorflow.keras.preprocessing import sequence\n",
        "from tcn import TCN"
      ],
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SUNkUL36Fijy",
        "colab_type": "text"
      },
      "source": [
        "### Define the max features"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cEIAVRBCEszU",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "max_features = 20000"
      ],
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FbXV0CnvE6LR",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "maxlen = 100\n",
        "batch_size = 32"
      ],
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qqC5lBXAFpMn",
        "colab_type": "text"
      },
      "source": [
        "### Load the IMDB Dataset"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-Wm-VCsuE_IH",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "0c38b768-213e-4885-cc44-5fc1786596d0"
      },
      "source": [
        "print(\"Loading Data...\")\n",
        "(x_train, y_train), (x_test, y_test) = imdb.load_data(num_words=max_features)"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Loading Data...\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZiMc4a8KFuhI",
        "colab_type": "text"
      },
      "source": [
        "### Print the shape of dataset"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GgQUVECXFLvx",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 85
        },
        "outputId": "c7038f47-029a-4d78-b006-221b41b04cb3"
      },
      "source": [
        "print(f\"X_train shape: {x_train.shape}\")\n",
        "print(f\"y_train shape: {y_train.shape}\")\n",
        "print(f\"X_test shape: {x_test.shape}\")\n",
        "print(f\"y_test shape: {y_test.shape}\")"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "X_train shape: (25000,)\n",
            "y_train shape: (25000,)\n",
            "X_test shape: (25000,)\n",
            "y_test shape: (25000,)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Vur0ppy0FxpO",
        "colab_type": "text"
      },
      "source": [
        "### Pad Sequence(Samples X Time)\n",
        "* Pad Sequence added and see the shape for X "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KZ9Q8iouFaAq",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        },
        "outputId": "ad6121e1-8ebd-4660-d12b-0130671b653b"
      },
      "source": [
        "x_train = sequence.pad_sequences(x_train, maxlen = maxlen)\n",
        "print(f\"X_train shape: {x_train.shape}\")\n",
        "x_test = sequence.pad_sequences(x_test, maxlen = maxlen)\n",
        "print(f\"X_test shape: {x_test.shape}\")"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "X_train shape: (25000, 100)\n",
            "X_test shape: (25000, 100)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SZw8xthoGaUO",
        "colab_type": "text"
      },
      "source": [
        "* Convert target y into Numpy Array "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "T8mlMSpdGE1H",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        },
        "outputId": "7a61f516-7fa3-4e0c-d9a8-613b0c7dc072"
      },
      "source": [
        "y_train = np.array(y_train)\n",
        "print(f\"y_train shape: {y_train.shape}\")\n",
        "y_test = np.array(y_test)\n",
        "print(f\"y_test shape: {y_test.shape}\")"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "y_train shape: (25000,)\n",
            "y_test shape: (25000,)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "N-jsYB_aGh1D",
        "colab_type": "text"
      },
      "source": [
        "### Design the TCN Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rokkaoJkGPxX",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 323
        },
        "outputId": "b8cf064e-17ea-4ed1-b415-eb3321e6e81b"
      },
      "source": [
        "i = Input(shape = (maxlen,))\n",
        "x = Embedding(max_features, 128)(i)\n",
        "x = TCN(nb_filters = 64, kernel_size=6, dilations=[1,2,4,8,16,32,64])(x)\n",
        "x = Dropout(0.5)(x)\n",
        "x = Dense(1, activation = 'sigmoid')(x)\n",
        "model = Model(inputs=[i], outputs=[x])\n",
        "model.compile(optimizer='sgd', loss='binary_crossentropy', metrics = ['accuracy'])\n",
        "model.summary()"
      ],
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"functional_3\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "input_2 (InputLayer)         [(None, 100)]             0         \n",
            "_________________________________________________________________\n",
            "embedding_1 (Embedding)      (None, 100, 128)          2560000   \n",
            "_________________________________________________________________\n",
            "tcn_1 (TCN)                  (None, 64)                377792    \n",
            "_________________________________________________________________\n",
            "dropout_1 (Dropout)          (None, 64)                0         \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 1)                 65        \n",
            "=================================================================\n",
            "Total params: 2,937,857\n",
            "Trainable params: 2,937,857\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mEfetpDtHtIF",
        "colab_type": "text"
      },
      "source": [
        "### Model Training"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Q56zIcSwHLku",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 357
        },
        "outputId": "5fef6b58-91ce-40ea-edac-6034c4422635"
      },
      "source": [
        "history = model.fit(x_train, y_train, batch_size = batch_size, epochs = 10, validation_data = [x_test, y_test])"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/10\n",
            "782/782 [==============================] - 13s 17ms/step - loss: 0.6969 - accuracy: 0.5008 - val_loss: 0.0000e+00 - val_accuracy: 0.0000e+00\n",
            "Epoch 2/10\n",
            "782/782 [==============================] - 13s 16ms/step - loss: 0.6910 - accuracy: 0.5264 - val_loss: 0.0000e+00 - val_accuracy: 0.0000e+00\n",
            "Epoch 3/10\n",
            "782/782 [==============================] - 13s 16ms/step - loss: 0.6892 - accuracy: 0.5370 - val_loss: 0.0000e+00 - val_accuracy: 0.0000e+00\n",
            "Epoch 4/10\n",
            "782/782 [==============================] - 13s 16ms/step - loss: 0.6869 - accuracy: 0.5500 - val_loss: 0.0000e+00 - val_accuracy: 0.0000e+00\n",
            "Epoch 5/10\n",
            "782/782 [==============================] - 13s 16ms/step - loss: 0.6838 - accuracy: 0.5562 - val_loss: 0.0000e+00 - val_accuracy: 0.0000e+00\n",
            "Epoch 6/10\n",
            "782/782 [==============================] - 13s 16ms/step - loss: 0.6737 - accuracy: 0.5838 - val_loss: 0.0000e+00 - val_accuracy: 0.0000e+00\n",
            "Epoch 7/10\n",
            "782/782 [==============================] - 13s 16ms/step - loss: 0.6542 - accuracy: 0.6164 - val_loss: 0.0000e+00 - val_accuracy: 0.0000e+00\n",
            "Epoch 8/10\n",
            "782/782 [==============================] - 13s 16ms/step - loss: 0.6248 - accuracy: 0.6509 - val_loss: 0.0000e+00 - val_accuracy: 0.0000e+00\n",
            "Epoch 9/10\n",
            "782/782 [==============================] - 13s 16ms/step - loss: 0.5792 - accuracy: 0.6982 - val_loss: 0.0000e+00 - val_accuracy: 0.0000e+00\n",
            "Epoch 10/10\n",
            "782/782 [==============================] - 13s 16ms/step - loss: 0.5362 - accuracy: 0.7342 - val_loss: 0.0000e+00 - val_accuracy: 0.0000e+00\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "J4hDj23RIIcD",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        },
        "outputId": "5a2e4cf2-9dc4-406a-8742-71560e91870b"
      },
      "source": [
        "result = model.evaluate(x_test, y_test)\n",
        "print(f\"Accuracy : {result[1] * 100:.2f} %\")\n",
        "print(f\"Loss : {result[0] * 100:.2f} %\")"
      ],
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "782/782 [==============================] - 4s 5ms/step - loss: 0.5595 - accuracy: 0.6995\n",
            "Accuracy : 69.95 %\n",
            "Loss : 55.95 %\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "H_pXfJLEINbT",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 533
        },
        "outputId": "d0ba2c31-217c-4350-f37a-5f28431675ad"
      },
      "source": [
        "from tensorflow import keras\n",
        "keras.utils.plot_model(model, show_shapes=True)"
      ],
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 20
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OSyBk3a2KACY",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "outputId": "5870baac-06a6-4535-e72f-75ca9fd5fa4b"
      },
      "source": [
        "from matplotlib import pyplot as plt\n",
        "%matplotlib inline\n",
        "plt.plot(history.history['accuracy'])\n",
        "plt.plot(history.history['val_accuracy'])\n",
        "plt.title('model accuracy')\n",
        "plt.ylabel('accuracy')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['train', 'val'], loc='upper left')\n",
        "plt.show()"
      ],
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HqnNi_VDOR9s",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "y_pred = model.predict(x_test)"
      ],
      "execution_count": 23,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UmhziU3pPLD5",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.metrics import accuracy_score\n",
        "from sklearn.metrics import precision_score\n",
        "from sklearn.metrics import recall_score\n",
        "from sklearn.metrics import f1_score\n",
        "from sklearn.metrics import cohen_kappa_score\n",
        "from sklearn.metrics import roc_auc_score\n",
        "from sklearn.metrics import confusion_matrix, classification_report\n",
        "import seaborn as sns"
      ],
      "execution_count": 39,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lKXBrZqAQTk4",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        },
        "outputId": "af09076e-182f-47cb-ea1c-1ef5d46521d4"
      },
      "source": [
        "yhat_probs = model.predict(x_test, verbose=0)\n",
        "print(y_test)\n",
        "print(np.argmax(yhat_probs, axis=1))"
      ],
      "execution_count": 32,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[0 1 1 ... 0 0 0]\n",
            "[0 0 0 ... 0 0 0]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7wUO-WlCPxOg",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 523
        },
        "outputId": "dc5a5842-9c8c-402c-cd08-1686c8b6a47b"
      },
      "source": [
        "# predict probabilities for test set\n",
        "yhat_probs = np.argmax(model.predict(x_test, verbose=0), axis=1)\n",
        " \n",
        "# kappa\n",
        "kappa = cohen_kappa_score(y_test, yhat_probs)\n",
        "print('Cohens kappa: %f' % kappa)\n",
        "# ROC AUC\n",
        "auc = roc_auc_score(y_test, yhat_probs)\n",
        "print('ROC AUC: %f' % auc)\n",
        "\n",
        "print(classification_report(y_test, yhat_probs))\n",
        "# confusion matrix\n",
        "matrix = confusion_matrix(y_test, yhat_probs)\n",
        "sns.heatmap(matrix, annot=True)\n"
      ],
      "execution_count": 41,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Cohens kappa: 0.000000\n",
            "ROC AUC: 0.500000\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.50      1.00      0.67     12500\n",
            "           1       0.00      0.00      0.00     12500\n",
            "\n",
            "    accuracy                           0.50     25000\n",
            "   macro avg       0.25      0.50      0.33     25000\n",
            "weighted avg       0.25      0.50      0.33     25000\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/sklearn/metrics/_classification.py:1272: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
            "  _warn_prf(average, modifier, msg_start, len(result))\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f0bc98e2c50>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 41
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}