raw_txt_to_word_emb.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "\n",
    "imdb_dir = '/home/joel/DL_Python/dl_text_and_seq/aclImdb'\n",
    "train_dir = os.path.join(imdb_dir, 'train')\n",
    "\n",
    "labels = []\n",
    "texts = []\n",
    "\n",
    "for label_type in ['neg', 'pos']:\n",
    "    dir_name = os.path.join(train_dir, label_type)\n",
    "    for fname in os.listdir(dir_name):\n",
    "        if fname[-4:] == '.txt':\n",
    "            f = open(os.path.join(dir_name, fname))\n",
    "            texts.append(f.read())\n",
    "            f.close()\n",
    "            if label_type == 'neg':\n",
    "                labels.append(0)\n",
    "            else:\n",
    "                labels.append(1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 88582 unique tokens.\n",
      "Shape of data tensor: (25000, 100)\n",
      "Shape of label tensor: (25000,)\n"
     ]
    }
   ],
   "source": [
    "from keras.preprocessing.text import Tokenizer\n",
    "from keras.preprocessing.sequence import pad_sequences\n",
    "import numpy as np\n",
    "\n",
    "maxlen = 100  # We will cut reviews after 100 words\n",
    "training_samples = 20000  # We will be training on 200 samples\n",
    "validation_samples = 10000  # We will be validating on 10000 samples\n",
    "max_words = 10000  # We will only consider the top 10,000 words in the dataset\n",
    "\n",
    "tokenizer = Tokenizer(num_words=max_words)\n",
    "tokenizer.fit_on_texts(texts)\n",
    "sequences = tokenizer.texts_to_sequences(texts)\n",
    "\n",
    "word_index = tokenizer.word_index\n",
    "print('Found %s unique tokens.' % len(word_index))\n",
    "\n",
    "data = pad_sequences(sequences, maxlen=maxlen)\n",
    "\n",
    "labels = np.asarray(labels)\n",
    "print('Shape of data tensor:', data.shape)\n",
    "print('Shape of label tensor:', labels.shape)\n",
    "\n",
    "# Split the data into a training set and a validation set\n",
    "# But first, shuffle the data, since we started from data\n",
    "# where sample are ordered (all negative first, then all positive).\n",
    "indices = np.arange(data.shape[0])\n",
    "np.random.shuffle(indices)\n",
    "data = data[indices]\n",
    "labels = labels[indices]\n",
    "\n",
    "x_train = data[:training_samples]\n",
    "y_train = labels[:training_samples]\n",
    "x_val = data[training_samples: training_samples + validation_samples]\n",
    "y_val = labels[training_samples: training_samples + validation_samples]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 400000 word vectors.\n"
     ]
    }
   ],
   "source": [
    "glove_dir = '/home/joel/Downloads/glove.6B'\n",
    "\n",
    "embeddings_index = {}\n",
    "f = open(os.path.join(glove_dir, 'glove.6B.100d.txt'))\n",
    "for line in f:\n",
    "    values = line.split()\n",
    "    word = values[0]\n",
    "    coefs = np.asarray(values[1:], dtype='float32')\n",
    "    embeddings_index[word] = coefs\n",
    "f.close()\n",
    "\n",
    "print('Found %s word vectors.' % len(embeddings_index))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "embedding_dim = 100\n",
    "\n",
    "embedding_matrix = np.zeros((max_words, embedding_dim))\n",
    "for word, i in word_index.items():\n",
    "    embedding_vector = embeddings_index.get(word)\n",
    "    if i < max_words:\n",
    "        if embedding_vector is not None:\n",
    "            # Words not found in embedding index will be all-zeros.\n",
    "            embedding_matrix[i] = embedding_vector"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From /home/joel/anaconda3/lib/python3.7/site-packages/tensorflow/python/framework/op_def_library.py:263: colocate_with (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Colocations handled automatically by placer.\n",
      "WARNING:tensorflow:From /home/joel/anaconda3/lib/python3.7/site-packages/keras/backend/tensorflow_backend.py:1205: calling reduce_prod_v1 (from tensorflow.python.ops.math_ops) with keep_dims is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "keep_dims is deprecated, use keepdims instead\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "embedding_1 (Embedding)      (None, 100, 100)          1000000   \n",
      "_________________________________________________________________\n",
      "flatten_1 (Flatten)          (None, 10000)             0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 32)                320032    \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 1)                 33        \n",
      "=================================================================\n",
      "Total params: 1,320,065\n",
      "Trainable params: 1,320,065\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import Embedding, Flatten, Dense\n",
    "\n",
    "model = Sequential()\n",
    "model.add(Embedding(max_words, embedding_dim, input_length=maxlen))\n",
    "model.add(Flatten())\n",
    "model.add(Dense(32, activation='relu'))\n",
    "model.add(Dense(1, activation='sigmoid'))\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.layers[0].set_weights([embedding_matrix])\n",
    "model.layers[0].trainable = False"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From /home/joel/anaconda3/lib/python3.7/site-packages/tensorflow/python/ops/math_ops.py:3066: to_int32 (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use tf.cast instead.\n",
      "Train on 200 samples, validate on 10000 samples\n",
      "Epoch 1/10\n",
      "200/200 [==============================] - 0s - loss: 1.7844 - acc: 0.5550 - val_loss: 0.7062 - val_acc: 0.5121\n",
      "Epoch 2/10\n",
      "200/200 [==============================] - 0s - loss: 0.5299 - acc: 0.7800 - val_loss: 0.6924 - val_acc: 0.5550\n",
      "Epoch 3/10\n",
      "200/200 [==============================] - 0s - loss: 0.4168 - acc: 0.7950 - val_loss: 0.6850 - val_acc: 0.5715\n",
      "Epoch 4/10\n",
      "200/200 [==============================] - 0s - loss: 0.3684 - acc: 0.7950 - val_loss: 0.6825 - val_acc: 0.5764\n",
      "Epoch 5/10\n",
      "200/200 [==============================] - 0s - loss: 0.1918 - acc: 0.9550 - val_loss: 0.7629 - val_acc: 0.5446\n",
      "Epoch 6/10\n",
      "200/200 [==============================] - 0s - loss: 0.1454 - acc: 0.9750 - val_loss: 1.7895 - val_acc: 0.5012\n",
      "Epoch 7/10\n",
      "200/200 [==============================] - 0s - loss: 0.1731 - acc: 0.9300 - val_loss: 0.8312 - val_acc: 0.5345\n",
      "Epoch 8/10\n",
      "200/200 [==============================] - 0s - loss: 0.0667 - acc: 1.0000 - val_loss: 0.7408 - val_acc: 0.5751\n",
      "Epoch 9/10\n",
      "200/200 [==============================] - 0s - loss: 0.0313 - acc: 1.0000 - val_loss: 0.8572 - val_acc: 0.5630\n",
      "Epoch 10/10\n",
      "200/200 [==============================] - 0s - loss: 0.0185 - acc: 1.0000 - val_loss: 0.8340 - val_acc: 0.5675\n"
     ]
    }
   ],
   "source": [
    "model.compile(optimizer='rmsprop',\n",
    "              loss='binary_crossentropy',\n",
    "              metrics=['acc'])\n",
    "history = model.fit(x_train, y_train,\n",
    "                    epochs=10,\n",
    "                    batch_size=32,\n",
    "                    validation_data=(x_val, y_val))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "import h5py\n",
    "model.save_weights('pre_trained_glove_model.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "acc = history.history['acc']\n",
    "val_acc = history.history['val_acc']\n",
    "loss = history.history['loss']\n",
    "val_loss = history.history['val_loss']\n",
    "\n",
    "epochs = range(1, len(acc) + 1)\n",
    "\n",
    "plt.plot(epochs, acc, 'bo', label='Training acc')\n",
    "plt.plot(epochs, val_acc, 'b', label='Validation acc')\n",
    "plt.title('Training and validation accuracy')\n",
    "plt.legend()\n",
    "\n",
    "plt.figure()\n",
    "\n",
    "plt.plot(epochs, loss, 'bo', label='Training loss')\n",
    "plt.plot(epochs, val_loss, 'b', label='Validation loss')\n",
    "plt.title('Training and validation loss')\n",
    "plt.legend()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "embedding_3 (Embedding)      (None, 100, 100)          1000000   \n",
      "_________________________________________________________________\n",
      "flatten_3 (Flatten)          (None, 10000)             0         \n",
      "_________________________________________________________________\n",
      "dense_5 (Dense)              (None, 32)                320032    \n",
      "_________________________________________________________________\n",
      "dense_6 (Dense)              (None, 1)                 33        \n",
      "=================================================================\n",
      "Total params: 1,320,065\n",
      "Trainable params: 1,320,065\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n",
      "Train on 20000 samples, validate on 5000 samples\n",
      "Epoch 1/10\n",
      "20000/20000 [==============================] - 16s - loss: 0.6939 - acc: 0.5014 - val_loss: 0.6946 - val_acc: 0.4846\n",
      "Epoch 2/10\n",
      "20000/20000 [==============================] - 18s - loss: 0.5018 - acc: 0.7583 - val_loss: 0.8926 - val_acc: 0.4918\n",
      "Epoch 3/10\n",
      "20000/20000 [==============================] - 18s - loss: 0.1206 - acc: 0.9575 - val_loss: 1.4490 - val_acc: 0.4824\n",
      "Epoch 4/10\n",
      "20000/20000 [==============================] - 17s - loss: 0.0294 - acc: 0.9931 - val_loss: 1.8759 - val_acc: 0.4882\n",
      "Epoch 5/10\n",
      "20000/20000 [==============================] - 15s - loss: 0.0128 - acc: 0.9974 - val_loss: 2.3276 - val_acc: 0.4842\n",
      "Epoch 6/10\n",
      "20000/20000 [==============================] - 16s - loss: 0.0065 - acc: 0.9977 - val_loss: 2.8664 - val_acc: 0.4854\n",
      "Epoch 7/10\n",
      "20000/20000 [==============================] - 15s - loss: 0.0080 - acc: 0.9976 - val_loss: 3.0582 - val_acc: 0.4834\n",
      "Epoch 8/10\n",
      "20000/20000 [==============================] - 15s - loss: 0.0057 - acc: 0.9981 - val_loss: 3.4759 - val_acc: 0.4790\n",
      "Epoch 9/10\n",
      "20000/20000 [==============================] - 16s - loss: 0.0052 - acc: 0.9982 - val_loss: 3.7008 - val_acc: 0.4786\n",
      "Epoch 10/10\n",
      "20000/20000 [==============================] - 18s - loss: 0.0048 - acc: 0.9981 - val_loss: 3.8000 - val_acc: 0.4806\n"
     ]
    }
   ],
   "source": [
    "#Training the same model without pretrained word embeddings\n",
    "\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Embedding, Flatten, Dense\n",
    "\n",
    "model = Sequential()\n",
    "model.add(Embedding(max_words, embedding_dim, input_length=maxlen))\n",
    "model.add(Flatten())\n",
    "model.add(Dense(32, activation='relu'))\n",
    "model.add(Dense(1, activation='sigmoid'))\n",
    "model.summary()\n",
    "\n",
    "model.compile(optimizer='rmsprop',\n",
    "              loss='binary_crossentropy',\n",
    "              metrics=['acc'])\n",
    "history = model.fit(x_train, y_train,\n",
    "                    epochs=10,\n",
    "                    batch_size=32,\n",
    "                    validation_data=(x_val, y_val))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "acc = history.history['acc']\n",
    "val_acc = history.history['val_acc']\n",
    "loss = history.history['loss']\n",
    "val_loss = history.history['val_loss']\n",
    "\n",
    "epochs = range(1, len(acc) + 1)\n",
    "\n",
    "plt.plot(epochs, acc, 'bo', label='Training acc')\n",
    "plt.plot(epochs, val_acc, 'b', label='Validation acc')\n",
    "plt.title('Training and validation accuracy')\n",
    "plt.legend()\n",
    "\n",
    "plt.figure()\n",
    "\n",
    "plt.plot(epochs, loss, 'bo', label='Training loss')\n",
    "plt.plot(epochs, val_loss, 'b', label='Validation loss')\n",
    "plt.title('Training and validation loss')\n",
    "plt.legend()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Evaluation of the model on the test data\n",
    "\n",
    "test_dir = os.path.join(imdb_dir, 'test')\n",
    "\n",
    "labels = []\n",
    "texts = []\n",
    "\n",
    "for label_type in ['neg', 'pos']:\n",
    "    dir_name = os.path.join(test_dir, label_type)\n",
    "    for fname in sorted(os.listdir(dir_name)):\n",
    "        if fname[-4:] == '.txt':\n",
    "            f = open(os.path.join(dir_name, fname))\n",
    "            texts.append(f.read())\n",
    "            f.close()\n",
    "            if label_type == 'neg':\n",
    "                labels.append(0)\n",
    "            else:\n",
    "                labels.append(1)\n",
    "\n",
    "sequences = tokenizer.texts_to_sequences(texts)\n",
    "x_test = pad_sequences(sequences, maxlen=maxlen)\n",
    "y_test = np.asarray(labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "24448/25000 [============================>.] - ETA: 0s"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[0.8339632079315186, 0.56128]"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.load_weights('pre_trained_glove_model.h5')\n",
    "model.evaluate(x_test, y_test)"
   ]
  }
 ],
 "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": "3.7.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}