toxic_conv_gru.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import csv\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from keras.models import load_model, Sequential\n",
    "from keras.layers import BatchNormalization, Conv1D, Dense, Dropout, Embedding, Flatten, GRU, MaxPooling1D\n",
    "from keras.preprocessing.text import Tokenizer\n",
    "from keras.preprocessing.sequence import pad_sequences \n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "from toxic_classifier_helper import ToxicClassifierHelper"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "VOCAB_SIZE = 25000\n",
    "MAX_TEXT_LEN = 100\n",
    "EMBEDDING_DIM = 128"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from importlib import reload\n",
    "import toxic_classifier_helper\n",
    "reload(toxic_classifier_helper)\n",
    "from toxic_classifier_helper import ToxicClassifierHelper"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "helper = ToxicClassifierHelper(vocab_size=VOCAB_SIZE, max_text_len=MAX_TEXT_LEN)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = helper.create_train_test_split(.98)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "model = Sequential()\n",
    "\n",
    "embedding = Embedding(VOCAB_SIZE, EMBEDDING_DIM)\n",
    "\n",
    "model.add(embedding)\n",
    "model.add(Dropout(.1))\n",
    "model.add(Conv1D(filters=32, kernel_size=2, padding=\"same\", activation=\"relu\"))\n",
    "model.add(BatchNormalization())\n",
    "model.add(Dropout(.1))\n",
    "model.add(MaxPooling1D(pool_size=2))\n",
    "model.add(Conv1D(filters=32, kernel_size=2, padding=\"same\", activation=\"relu\"))\n",
    "model.add(BatchNormalization())\n",
    "model.add(MaxPooling1D(pool_size=2))\n",
    "model.add(GRU(32))\n",
    "model.add(Dropout(.1))\n",
    "model.add(Dense(256, activation=\"relu\"))\n",
    "model.add(BatchNormalization())\n",
    "model.add(Dense(6, activation=\"sigmoid\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# fasttext_weights = np.load(\"fasttext/fasttext_weights.npy\")\n",
    "# model.layers[0].set_weights(fasttext_weights)\n",
    "# model.layers[0].trainable = False"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "model.compile(optimizer=\"adam\", loss=\"binary_crossentropy\", metrics=[\"accuracy\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "model.fit(x=np.array(X_train), y=np.array(y_train), validation_data=(X_test, y_test), epochs=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "helper.save_model(model, \"test.h5\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "model = load_model(\"saved_models/test.h5\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "submission_input = pd.read_csv(\"test_csv\")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "helper.create_submission(model, \"test_submission\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "dl",
   "language": "python",
   "name": "dl"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}