a_deeper_model.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Importing Dependencies"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense\n",
    "from keras.layers import Dropout\n",
    "from keras.layers import LSTM\n",
    "from keras.layers import RNN\n",
    "from keras.utils import np_utils\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Loading of Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "text = (open(\"/Users/pranjal/Desktop/text_generator/sonnets.txt\").read())\n",
    "text=text.lower()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Creating character/word mappings"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "characters = sorted(list(set(text)))\n",
    "\n",
    "n_to_char = {n:char for n, char in enumerate(characters)}\n",
    "char_to_n = {char:n for n, char in enumerate(characters)}\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data pre-processing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X = []\n",
    "Y = []\n",
    "length = len(text)\n",
    "seq_length = 100\n",
    "\n",
    "for i in range(0, length-seq_length, 1):\n",
    "    sequence = text[i:i + seq_length]\n",
    "    label =text[i + seq_length]\n",
    "    X.append([char_to_n[char] for char in sequence])\n",
    "    Y.append(char_to_n[label])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "X_modified = np.reshape(X, (len(X), seq_length, 1))\n",
    "X_modified = X_modified / float(len(characters))\n",
    "Y_modified = np_utils.to_categorical(Y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## A deeper model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "model = Sequential()\n",
    "model.add(LSTM(400, input_shape=(X_modified.shape[1], X_modified.shape[2]), return_sequences=True))\n",
    "model.add(Dropout(0.2))\n",
    "model.add(LSTM(400, return_sequences=True))\n",
    "model.add(Dropout(0.2))\n",
    "model.add(LSTM(400))\n",
    "model.add(Dropout(0.2))\n",
    "model.add(Dense(Y_modified.shape[1], activation='softmax'))\n",
    "\n",
    "model.compile(loss='categorical_crossentropy', optimizer='adam')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "model.fit(X_modified, Y_modified, epochs=100, batch_size=50)\n",
    "\n",
    "model.save_weights('text_generator_400_0.2_400_0.2_400_0.2_100.h5')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "model.load_weights('/Users/pranjal/Desktop/text_generator/models/text_generator_400_0.2_400_0.2_400_0.2_100.h5')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Generating Text"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "string_mapped = X[99]\n",
    "full_string = [n_to_char[value] for value in string_mapped]\n",
    "# generating characters\n",
    "for i in range(400):\n",
    "    x = np.reshape(string_mapped,(1,len(string_mapped), 1))\n",
    "    x = x / float(len(characters))\n",
    "\n",
    "    pred_index = np.argmax(model.predict(x, verbose=0))\n",
    "    seq = [n_to_char[value] for value in string_mapped]\n",
    "    full_string.append(n_to_char[pred_index])\n",
    "\n",
    "    string_mapped.append(pred_index)\n",
    "    string_mapped = string_mapped[1:len(string_mapped)]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\"s the riper should by time decease,\\n his tender heir might bear his memory:\\n but thou, contracted to the world's false sporoe,\\n with eyes so dond touls be thy domfornds,\\n which for memorion of the seasons new;\\n mike own self-love to shou art constant\\n how can i then be oy love doth give,\\n the rose looks fair, but fairer bomments age.\\n now with the seas that i have seen dychl\\n that mot to teed of you, if h ho bontent,\\n and all the ress of the heartai thee;\\n the world will was iis bound and farty \""
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#combining text\n",
    "txt=\"\"\n",
    "for char in full_string:\n",
    "    txt = txt+char\n",
    "txt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}