IMDB-RNN.ipynb

imdb-rnn.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "IMDB-RNN.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyMukNpmLTFyaT3ileqYbQgC",
      "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/gist/Melihemin/bcbac2083ebb05a99eef99c7c175c162/imdb-rnn.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DsDQOKYQHAeG"
      },
      "source": [
        "from keras.layers import SimpleRNN"
      ],
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nDfgFUdEJRQL"
      },
      "source": [
        "from keras.models import Sequential\n",
        "from keras.layers import Embedding"
      ],
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "93VuatkDWsIO"
      },
      "source": [
        "# Basit RNN"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wOtvpAqGJZTi"
      },
      "source": [
        "### Bir RNN Katmanı "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "kV3_wnurJRNd",
        "outputId": "f9a4a14e-63d3-4217-d950-c96f393a9da0"
      },
      "source": [
        "model = Sequential()\n",
        "model.add(Embedding(1000,32))\n",
        "model.add(SimpleRNN(32))\n",
        "model.summary()"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "embedding (Embedding)        (None, None, 32)          32000     \n",
            "_________________________________________________________________\n",
            "simple_rnn (SimpleRNN)       (None, 32)                2080      \n",
            "=================================================================\n",
            "Total params: 34,080\n",
            "Trainable params: 34,080\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WDmgkrsBJuCo"
      },
      "source": [
        "### Boyutlandırılmış RNN Katmanı"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "q8pnRq6VJRKa",
        "outputId": "496439f5-fab9-4bbb-c3a8-e15b6368daf6"
      },
      "source": [
        "model = Sequential()\n",
        "model.add(Embedding(1000,32))\n",
        "model.add(SimpleRNN(32, return_sequences=True))\n",
        "model.summary()"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_1\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "embedding_1 (Embedding)      (None, None, 32)          32000     \n",
            "_________________________________________________________________\n",
            "simple_rnn_1 (SimpleRNN)     (None, None, 32)          2080      \n",
            "=================================================================\n",
            "Total params: 34,080\n",
            "Trainable params: 34,080\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Wsnm0KKwJ-Bq"
      },
      "source": [
        "### Ardışık RNN Katmanı"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "d41pMMN8JRHc",
        "outputId": "0d9d4055-b4b7-4192-814e-aa462e060abf"
      },
      "source": [
        "model = Sequential()\n",
        "model.add(Embedding(1000,32))\n",
        "model.add(SimpleRNN(32, return_sequences=True))\n",
        "model.add(SimpleRNN(32, return_sequences=True))\n",
        "model.add(SimpleRNN(32, return_sequences=True))\n",
        "model.add(SimpleRNN(32))\n",
        "model.summary()"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_2\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "embedding_2 (Embedding)      (None, None, 32)          32000     \n",
            "_________________________________________________________________\n",
            "simple_rnn_2 (SimpleRNN)     (None, None, 32)          2080      \n",
            "_________________________________________________________________\n",
            "simple_rnn_3 (SimpleRNN)     (None, None, 32)          2080      \n",
            "_________________________________________________________________\n",
            "simple_rnn_4 (SimpleRNN)     (None, None, 32)          2080      \n",
            "_________________________________________________________________\n",
            "simple_rnn_5 (SimpleRNN)     (None, 32)                2080      \n",
            "=================================================================\n",
            "Total params: 40,320\n",
            "Trainable params: 40,320\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QiPKLYBJKHyd"
      },
      "source": [
        "# Veri Kümesini Hazırlamak"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WQOoUPdHJREz"
      },
      "source": [
        "from keras.datasets import imdb\n",
        "from keras.preprocessing import sequence"
      ],
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FSEbI0EjJRBJ"
      },
      "source": [
        "num_features = 1000\n",
        "maxlen = 500\n",
        "batch_size = 32"
      ],
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "im-s1UDHJQ-S",
        "outputId": "0009d58c-df7d-483c-f080-6bf047170d6a"
      },
      "source": [
        "print('Load data..')\n",
        "(input_train, y_train), (input_test, y_test) = imdb.load_data(num_words=num_features)\n",
        "\n",
        "print(len(input_train), 'Eğitim Dizisi')\n",
        "print(len(input_test), 'Test Dizisi')"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Load data..\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "<string>:6: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n",
            "/usr/local/lib/python3.7/dist-packages/tensorflow/python/keras/datasets/imdb.py:159: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n",
            "  x_train, y_train = np.array(xs[:idx]), np.array(labels[:idx])\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "25000 Eğitim Dizisi\n",
            "25000 Test Dizisi\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/tensorflow/python/keras/datasets/imdb.py:160: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray\n",
            "  x_test, y_test = np.array(xs[idx:]), np.array(labels[idx:])\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CHxvJucZMFbq"
      },
      "source": [
        "input_train = sequence.pad_sequences(input_train,maxlen = maxlen)\n",
        "input_test = sequence.pad_sequences(input_test,maxlen = maxlen)"
      ],
      "execution_count": 17,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6B75Xb87LEsv"
      },
      "source": [
        "### Eğitim"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pr9Fc6zMJQ7S"
      },
      "source": [
        "from keras.layers import Dense\n",
        "from keras import layers"
      ],
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-pDyzc7VJQ5B"
      },
      "source": [
        "model = Sequential()\n",
        "model.add(Embedding(num_features, 32))\n",
        "model.add(SimpleRNN(32))\n",
        "model.add(Dense(1, activation='sigmoid'))"
      ],
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4r69vBfKLIqy",
        "outputId": "15d9471b-a140-44bb-b608-93b4600312a8"
      },
      "source": [
        "model.compile(optimizer='rmsprop', loss='binary_crossentropy', metrics=['accuracy'])\n",
        "history = model.fit(input_train, y_train,\n",
        "                    epochs=20,\n",
        "                    batch_size = 128,\n",
        "                    validation_split = 0.2)\n"
      ],
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/20\n",
            "157/157 [==============================] - 23s 142ms/step - loss: 0.4141 - accuracy: 0.8233 - val_loss: 0.4322 - val_accuracy: 0.7998\n",
            "Epoch 2/20\n",
            "157/157 [==============================] - 22s 139ms/step - loss: 0.3716 - accuracy: 0.8427 - val_loss: 0.3745 - val_accuracy: 0.8512\n",
            "Epoch 3/20\n",
            "157/157 [==============================] - 22s 141ms/step - loss: 0.3551 - accuracy: 0.8514 - val_loss: 0.3589 - val_accuracy: 0.8518\n",
            "Epoch 4/20\n",
            "157/157 [==============================] - 22s 139ms/step - loss: 0.3397 - accuracy: 0.8587 - val_loss: 0.3923 - val_accuracy: 0.8458\n",
            "Epoch 5/20\n",
            "157/157 [==============================] - 22s 139ms/step - loss: 0.3291 - accuracy: 0.8641 - val_loss: 0.4110 - val_accuracy: 0.8440\n",
            "Epoch 6/20\n",
            "157/157 [==============================] - 22s 139ms/step - loss: 0.3228 - accuracy: 0.8680 - val_loss: 0.4545 - val_accuracy: 0.8052\n",
            "Epoch 7/20\n",
            "157/157 [==============================] - 22s 141ms/step - loss: 0.2972 - accuracy: 0.8816 - val_loss: 0.4296 - val_accuracy: 0.8252\n",
            "Epoch 8/20\n",
            "157/157 [==============================] - 22s 140ms/step - loss: 0.2979 - accuracy: 0.8787 - val_loss: 0.4982 - val_accuracy: 0.7634\n",
            "Epoch 9/20\n",
            "157/157 [==============================] - 22s 141ms/step - loss: 0.2636 - accuracy: 0.8938 - val_loss: 0.5670 - val_accuracy: 0.7442\n",
            "Epoch 10/20\n",
            "157/157 [==============================] - 22s 140ms/step - loss: 0.2481 - accuracy: 0.9010 - val_loss: 0.5764 - val_accuracy: 0.7742\n",
            "Epoch 11/20\n",
            "157/157 [==============================] - 22s 140ms/step - loss: 0.2405 - accuracy: 0.9050 - val_loss: 0.5405 - val_accuracy: 0.8164\n",
            "Epoch 12/20\n",
            "157/157 [==============================] - 22s 139ms/step - loss: 0.2286 - accuracy: 0.9140 - val_loss: 0.4938 - val_accuracy: 0.8222\n",
            "Epoch 13/20\n",
            "157/157 [==============================] - 22s 139ms/step - loss: 0.2002 - accuracy: 0.9255 - val_loss: 0.5096 - val_accuracy: 0.7962\n",
            "Epoch 14/20\n",
            "157/157 [==============================] - 22s 140ms/step - loss: 0.1960 - accuracy: 0.9269 - val_loss: 0.5300 - val_accuracy: 0.8182\n",
            "Epoch 15/20\n",
            "157/157 [==============================] - 22s 139ms/step - loss: 0.1745 - accuracy: 0.9334 - val_loss: 0.5706 - val_accuracy: 0.8172\n",
            "Epoch 16/20\n",
            "157/157 [==============================] - 22s 139ms/step - loss: 0.1648 - accuracy: 0.9409 - val_loss: 0.5706 - val_accuracy: 0.8102\n",
            "Epoch 17/20\n",
            "157/157 [==============================] - 22s 140ms/step - loss: 0.1536 - accuracy: 0.9442 - val_loss: 0.5867 - val_accuracy: 0.8158\n",
            "Epoch 18/20\n",
            "157/157 [==============================] - 22s 139ms/step - loss: 0.1661 - accuracy: 0.9409 - val_loss: 0.6057 - val_accuracy: 0.8162\n",
            "Epoch 19/20\n",
            "157/157 [==============================] - 22s 140ms/step - loss: 0.1322 - accuracy: 0.9531 - val_loss: 0.6432 - val_accuracy: 0.8068\n",
            "Epoch 20/20\n",
            "157/157 [==============================] - 22s 139ms/step - loss: 0.1243 - accuracy: 0.9568 - val_loss: 0.6879 - val_accuracy: 0.7824\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "d4StUlvpP-T9"
      },
      "source": [
        "### Sonuçlar"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PyxAEWkLLInJ"
      },
      "source": [
        "import matplotlib.pyplot as plt"
      ],
      "execution_count": 30,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 214
        },
        "id": "q_yo15TILIim",
        "outputId": "fcb1f8a8-b52c-4726-86be-7b75fa0063ca"
      },
      "source": [
        "acc = history.history['acc']\n",
        "val_acc = history.history['val_acc']\n",
        "loss = history.history['loss']\n",
        "val_loss = history.history['val_loss']"
      ],
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "error",
          "ename": "KeyError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mKeyError\u001b[0m                                  Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-27-0642ce207f8b>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0macc\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mhistory\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mhistory\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'acc'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      2\u001b[0m \u001b[0mval_acc\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mhistory\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mhistory\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'val_acc'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0mloss\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mhistory\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mhistory\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'loss'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0mval_loss\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mhistory\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mhistory\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;34m'val_loss'\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mKeyError\u001b[0m: 'acc'"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rqzPuG_cLIgW"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TAcaLF24Wwhf"
      },
      "source": [
        "# Basit LSTM"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Ru5Dv_gIWz_h"
      },
      "source": [
        "model = Sequential()\n",
        "model.add(layers.Embedding(num_features, 32))\n",
        "model.add(layers.Bidirectional(layers.LSTM(32)))\n",
        "model.add(Dense(1, activation='sigmoid'))"
      ],
      "execution_count": 31,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "LXoFHRDRXHsU",
        "outputId": "c4fb8dee-0857-42c1-e4ef-e6190069c114"
      },
      "source": [
        "model.compile(optimizer='rmsprop', loss='binary_crossentropy', metrics=['accuracy'])\n",
        "history = model.fit(input_train, y_train,\n",
        "                    epochs=20,\n",
        "                    batch_size = 128,\n",
        "                    validation_split = 0.2)\n"
      ],
      "execution_count": 32,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/20\n",
            "157/157 [==============================] - 100s 616ms/step - loss: 0.6424 - accuracy: 0.6203 - val_loss: 0.5494 - val_accuracy: 0.7126\n",
            "Epoch 2/20\n",
            "157/157 [==============================] - 95s 606ms/step - loss: 0.4201 - accuracy: 0.8186 - val_loss: 0.4197 - val_accuracy: 0.8084\n",
            "Epoch 3/20\n",
            "157/157 [==============================] - 95s 605ms/step - loss: 0.3713 - accuracy: 0.8496 - val_loss: 0.4407 - val_accuracy: 0.7930\n",
            "Epoch 4/20\n",
            "157/157 [==============================] - 95s 604ms/step - loss: 0.3448 - accuracy: 0.8580 - val_loss: 0.3847 - val_accuracy: 0.8300\n",
            "Epoch 5/20\n",
            "157/157 [==============================] - 95s 605ms/step - loss: 0.3399 - accuracy: 0.8628 - val_loss: 0.3619 - val_accuracy: 0.8414\n",
            "Epoch 6/20\n",
            "157/157 [==============================] - 95s 606ms/step - loss: 0.3151 - accuracy: 0.8718 - val_loss: 0.3900 - val_accuracy: 0.8292\n",
            "Epoch 7/20\n",
            "157/157 [==============================] - 95s 606ms/step - loss: 0.3169 - accuracy: 0.8708 - val_loss: 0.3521 - val_accuracy: 0.8530\n",
            "Epoch 8/20\n",
            "157/157 [==============================] - 95s 608ms/step - loss: 0.3016 - accuracy: 0.8763 - val_loss: 0.3549 - val_accuracy: 0.8514\n",
            "Epoch 9/20\n",
            "157/157 [==============================] - 97s 617ms/step - loss: 0.2991 - accuracy: 0.8796 - val_loss: 0.3441 - val_accuracy: 0.8540\n",
            "Epoch 10/20\n",
            "157/157 [==============================] - 99s 633ms/step - loss: 0.2991 - accuracy: 0.8771 - val_loss: 0.3581 - val_accuracy: 0.8540\n",
            "Epoch 11/20\n",
            "157/157 [==============================] - 97s 621ms/step - loss: 0.2933 - accuracy: 0.8824 - val_loss: 0.3929 - val_accuracy: 0.8408\n",
            "Epoch 12/20\n",
            "157/157 [==============================] - 96s 614ms/step - loss: 0.2950 - accuracy: 0.8779 - val_loss: 0.4227 - val_accuracy: 0.7998\n",
            "Epoch 13/20\n",
            "157/157 [==============================] - 96s 609ms/step - loss: 0.3015 - accuracy: 0.8775 - val_loss: 0.4458 - val_accuracy: 0.8408\n",
            "Epoch 14/20\n",
            "157/157 [==============================] - 97s 618ms/step - loss: 0.2915 - accuracy: 0.8807 - val_loss: 0.3431 - val_accuracy: 0.8596\n",
            "Epoch 15/20\n",
            "157/157 [==============================] - 95s 605ms/step - loss: 0.2771 - accuracy: 0.8857 - val_loss: 0.3763 - val_accuracy: 0.8410\n",
            "Epoch 16/20\n",
            "157/157 [==============================] - 95s 606ms/step - loss: 0.2830 - accuracy: 0.8831 - val_loss: 0.3446 - val_accuracy: 0.8538\n",
            "Epoch 17/20\n",
            "157/157 [==============================] - 95s 606ms/step - loss: 0.2739 - accuracy: 0.8881 - val_loss: 0.4817 - val_accuracy: 0.7978\n",
            "Epoch 18/20\n",
            "157/157 [==============================] - 95s 606ms/step - loss: 0.2727 - accuracy: 0.8873 - val_loss: 0.3416 - val_accuracy: 0.8614\n",
            "Epoch 19/20\n",
            "157/157 [==============================] - 95s 607ms/step - loss: 0.2639 - accuracy: 0.8927 - val_loss: 0.3607 - val_accuracy: 0.8526\n",
            "Epoch 20/20\n",
            "157/157 [==============================] - 95s 606ms/step - loss: 0.2570 - accuracy: 0.8956 - val_loss: 0.3788 - val_accuracy: 0.8418\n"
          ],
          "name": "stdout"
        }
      ]
    }
  ]
}