imdb_pretrained_model.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "75559b15",
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "import os"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e84ffa8a",
   "metadata": {},
   "source": [
    "# Processing the labels of the raw IMDB data\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "edc14d27",
   "metadata": {},
   "outputs": [],
   "source": [
    "#imdb_dir  = \"E:/QUARTER 3_COMPUTER_VISION/nlp chapeter #6/archive (2)/aclImdb\"\n",
    "#train_dir = \"E:/QUARTER 3_COMPUTER_VISION/nlp chapeter #6/archive (2)/aclImdb/train\"\n",
    "\n",
    "imdb_dir  = \"aclImdb\"\n",
    "train_dir = \"aclImdb/train\"\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':  #   optional\n",
    "            \n",
    "            f = open(os.path.join(dir_name, fname ) , encoding='utf-8')\n",
    "            texts.append(f.read())\n",
    "            f.close()\n",
    "            if label_type == 'neg':\n",
    "                \n",
    "                labels.append(0)\n",
    "                \n",
    "            else:\n",
    "                labels.append(1)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c7dde5da",
   "metadata": {},
   "source": [
    "# Tokenizing the text of the raw IMDB data\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "d32d61f6",
   "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 tensorflow.keras.preprocessing.text import Tokenizer\n",
    "from tensorflow.keras.preprocessing.sequence import pad_sequences\n",
    "import numpy as np\n",
    "maxlen = 100\n",
    "training_samples = 200\n",
    "validation_samples = 10000\n",
    "max_words = 10000\n",
    "\n",
    "tokenizer = Tokenizer(num_words=max_words)\n",
    "tokenizer.fit_on_texts(texts)\n",
    "sequences = tokenizer.texts_to_sequences(texts)\n",
    "word_index = tokenizer.word_index\n",
    "print('Found %s unique tokens.' % len(word_index))\n",
    "data = pad_sequences(sequences, maxlen=maxlen)\n",
    "labels = np.asarray(labels)\n",
    "print('Shape of data tensor:', data.shape)\n",
    "print('Shape of label tensor:', labels.shape)\n",
    "\n",
    "indices = np.arange(data.shape[0])\n",
    "np.random.shuffle(indices)\n",
    "data = data[indices]\n",
    "labels = labels[indices]\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": 44,
   "id": "51299a7e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[  10,  141,   27, ...,    9,   67,   76],\n",
       "       [  78,   21,  434, ..., 1041,  151, 1231],\n",
       "       [ 257,   32, 1585, ...,    7,  542, 1176],\n",
       "       ...,\n",
       "       [  23,   21,  711, ...,  404,  287,    9],\n",
       "       [  13,  458,    3, ...,   23,   52, 1127],\n",
       "       [ 120,   50,    4, ...,    9, 1910,   11]])"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "id": "e60ca214",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "100"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(data[0])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "ffa90feb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0, 0, 1, ..., 1, 1, 1])"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "labels\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "id": "b20b61e6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "200"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "training_samples\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d8865e6b",
   "metadata": {},
   "outputs": [],
   "source": [
    "glove_dir = '/Users/fchollet/Downloads/glove.6B'\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",
    "print('Found %s word vectors.' % len(embeddings_index))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c56058cd",
   "metadata": {},
   "source": [
    "# Parsing the GloVe word-embeddings file\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "id": "ba4658a8",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 400000 word vectors.\n"
     ]
    }
   ],
   "source": [
    "\n",
    "glove_data = \"glove/glove.6B.100d.txt\"\n",
    "embeddings_index = {}\n",
    "\n",
    "f = open(glove_data, encoding=\"utf8\" )\n",
    "\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",
    "print('Found %s word vectors.' % len(embeddings_index))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "24722345",
   "metadata": {},
   "outputs": [],
   "source": [
    "embedding_matrix = np.zeros((max_words, 100))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "65b2818d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(10000, 100)"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "embedding_matrix.shape\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "id": "a4fed2d2",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0., 0., 0., ..., 0., 0., 0.],\n",
       "       [0., 0., 0., ..., 0., 0., 0.],\n",
       "       [0., 0., 0., ..., 0., 0., 0.],\n",
       "       ...,\n",
       "       [0., 0., 0., ..., 0., 0., 0.],\n",
       "       [0., 0., 0., ..., 0., 0., 0.],\n",
       "       [0., 0., 0., ..., 0., 0., 0.]])"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "embedding_matrix\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "60204df2",
   "metadata": {},
   "source": [
    "# Preparing the GloVe word-embeddings matrix\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "283dcd38",
   "metadata": {},
   "outputs": [],
   "source": [
    "embedding_dim = 100\n",
    "embedding_matrix = np.zeros((max_words, embedding_dim))\n",
    "for word, i in word_index.items():\n",
    "    if i < max_words:        \n",
    "        embedding_vector = embeddings_index.get(word)\n",
    "        if embedding_vector is not None:\n",
    "            embedding_matrix[i] = embedding_vector"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "35f74807",
   "metadata": {},
   "source": [
    "# Model definition\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "a19cb673",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_1\"\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_2 (Dense)              (None, 32)                320032    \n",
      "_________________________________________________________________\n",
      "dense_3 (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 tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Embedding, Flatten, Dense\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": "markdown",
   "id": "0c951cff",
   "metadata": {},
   "source": [
    "# Loading pretrained word embeddings into the Embedding layer\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "f83ec8db",
   "metadata": {},
   "outputs": [],
   "source": [
    "model.layers[0].set_weights([embedding_matrix])\n",
    "model.layers[0].trainable = False"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d469a58a",
   "metadata": {},
   "source": [
    "# Training and evaluation\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "6b530fc3",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "7/7 [==============================] - 1s 187ms/step - loss: 2.1396 - acc: 0.5100 - val_loss: 0.6916 - val_acc: 0.5417\n",
      "Epoch 2/10\n",
      "7/7 [==============================] - 1s 121ms/step - loss: 0.5783 - acc: 0.7000 - val_loss: 1.3289 - val_acc: 0.4961\n",
      "Epoch 3/10\n",
      "7/7 [==============================] - 1s 117ms/step - loss: 0.5000 - acc: 0.8150 - val_loss: 0.7351 - val_acc: 0.5263\n",
      "Epoch 4/10\n",
      "7/7 [==============================] - 1s 121ms/step - loss: 0.3848 - acc: 0.8500 - val_loss: 0.8550 - val_acc: 0.5159\n",
      "Epoch 5/10\n",
      "7/7 [==============================] - 1s 120ms/step - loss: 0.2181 - acc: 0.9450 - val_loss: 0.9058 - val_acc: 0.5138\n",
      "Epoch 6/10\n",
      "7/7 [==============================] - 1s 117ms/step - loss: 0.3070 - acc: 0.8650 - val_loss: 0.8049 - val_acc: 0.5574\n",
      "Epoch 7/10\n",
      "7/7 [==============================] - 1s 123ms/step - loss: 0.0853 - acc: 0.9950 - val_loss: 0.7515 - val_acc: 0.5811\n",
      "Epoch 8/10\n",
      "7/7 [==============================] - 1s 120ms/step - loss: 0.0454 - acc: 1.0000 - val_loss: 1.0579 - val_acc: 0.5304\n",
      "Epoch 9/10\n",
      "7/7 [==============================] - 1s 119ms/step - loss: 0.0828 - acc: 0.9950 - val_loss: 1.1414 - val_acc: 0.5280\n",
      "Epoch 10/10\n",
      "7/7 [==============================] - 1s 119ms/step - loss: 0.0308 - acc: 1.0000 - val_loss: 0.7767 - val_acc: 0.5929\n"
     ]
    }
   ],
   "source": [
    "model.compile(optimizer='rmsprop', loss='binary_crossentropy', metrics=['acc'])\n",
    "\n",
    "history = model.fit(x_train, y_train, epochs=10, validation_data=(x_val, y_val))\n",
    "\n",
    "model.save_weights('pre_trained_glove_model.h5')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "394aa7ba",
   "metadata": {},
   "source": [
    "# Plotting the results\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "f9941c1f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAEICAYAAABPgw/pAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAAAsTAAALEwEAmpwYAAAos0lEQVR4nO3deXxU9b3/8deHRTAgCIKIRBJoQcQiASIKLkVFxWKhKv4EUgVppWrVqrdWWhfQyr3eSqv11uVSxRWK1irXtriiFLeqwRVQFJElrhEVkEUI+fz++E7CZMgyIZPM5OT9fDzmMXOWOeczZ5LPfM/3fM/3a+6OiIg0fs3SHYCIiKSGErqISEQooYuIRIQSuohIRCihi4hEhBK6iEhEKKFHmJk9ZmYTUr1uOpnZKjMbXg/bdTP7buz17WZ2VTLr7sZ+Cszsyd2NU6Q6pnbomcXMvombzAK+BXbEpn/m7rMbPqrMYWargJ+6+9Mp3q4Dvdx9RarWNbNc4EOgpbuXpCRQkWq0SHcAUpG7ty17XV3yMrMWShKSKfT3mBlU5dJImNkwMysys8vN7FPgLjPrYGb/MLNiM/sq9jo77j0LzeynsdcTzex5M5sRW/dDMztpN9ftYWaLzGyjmT1tZreY2f1VxJ1MjL81sxdi23vSzDrFLT/TzFab2Tozu6Ka43O4mX1qZs3j5p1iZm/FXg82s5fM7Gsz+8TM/mRme1SxrbvN7Lq46cti7/nYzCYlrDvSzF43sw1mttbMpsUtXhR7/trMvjGzIWXHNu79Q83sVTNbH3semuyxqeVx7mhmd8U+w1dmNi9u2WgzeyP2GT4wsxGx+RWqt8xsWtn3bGa5saqnn5jZGuCZ2Py/xr6H9bG/kYPj3r+nmf0+9n2uj/2N7Wlm/zSzCxM+z1tm9qPKPqtUTQm9cdkP6AjkAJMJ399dsenuwBbgT9W8/zBgOdAJ+B1wp5nZbqw7B3gF2AeYBpxZzT6TiXE8cDawL7AH8EsAM+sL3Bbb/v6x/WVTCXf/N7AJODZhu3Nir3cAl8Q+zxDgOOD8auImFsOIWDzHA72AxPr7TcBZwN7ASOC8uER0dOx5b3dv6+4vJWy7I/BP4ObYZ/sD8E8z2yfhM+xybCpR03G+j1CFd3BsWzfGYhgM3AtcFvsMRwOrqthHZb4PHAScGJt+jHCc9gVeA+KrCGcAg4ChhL/jXwGlwD3Aj8tWMrP+QDdgfi3iEAB31yNDH4R/rOGx18OAbUDratbPA76Km15IqLIBmAisiFuWBTiwX23WJSSLEiArbvn9wP1JfqbKYrwybvp84PHY66uBuXHL2sSOwfAqtn0dMCv2ei9Css2pYt2LgUfiph34buz13cB1sdezgOvj1usdv24l270JuDH2Oje2bou45ROB52OvzwReSXj/S8DEmo5NbY4z0JWQODtUst7/lsVb3d9fbHpa2fcc99l6VhPD3rF12hN+cLYA/StZrxXwJeG6BITEf2t9/E9F/aESeuNS7O5byybMLMvM/jd2CruBcIq/d3y1Q4JPy164++bYy7a1XHd/4Mu4eQBrqwo4yRg/jXu9OS6m/eO37e6bgHVV7YtQGj/VzFoBpwKvufvqWBy9Y9UQn8bi+E9Cab0mFWIAVid8vsPM7NlYVcd64Nwkt1u27dUJ81YTSqdlqjo2FdRwnA8gfGdfVfLWA4APkoy3MuXHxsyam9n1sWqbDews6XeKPVpXti93/xZ4EPixmTUDxhHOKKSWlNAbl8QmSf8BHAgc5u7t2HmKX1U1Sip8AnQ0s6y4eQdUs35dYvwkftuxfe5T1cruvoyQEE+iYnULhKqbdwmlwHbAb3YnBsIZSrw5wKPAAe7eHrg9brs1NSH7mFBFEq878FEScSWq7jivJXxne1fyvrXAd6rY5ibC2VmZ/SpZJ/4zjgdGE6ql2hNK8WUxfAFsrWZf9wAFhKqwzZ5QPSXJUUJv3PYinMZ+HauPnVrfO4yVeAuBaWa2h5kNAX5YTzE+BJxsZkfGLmBeS81/s3OAiwgJ7a8JcWwAvjGzPsB5ScbwIDDRzPrGflAS49+LUPrdGquPHh+3rJhQ1dGzim3PB3qb2Xgza2FmZwB9gX8kGVtiHJUeZ3f/hFC3fWvs4mlLMytL+HcCZ5vZcWbWzMy6xY4PwBvA2Nj6+cCYJGL4lnAWlUU4CyqLoZRQffUHM9s/VpofEjubIpbAS4Hfo9L5blNCb9xuAvYklH7+DTzeQPstIFxYXEeot36A8I9cmZvYzRjdfSnwc0KS/gT4Ciiq4W1/IVxveMbdv4ib/0tCst0I/DkWczIxPBb7DM8AK2LP8c4HrjWzjYQ6/wfj3rsZmA68YKF1zeEJ214HnEwoXa8jXCQ8OSHuZN1E9cf5TGA74Szlc8I1BNz9FcJF1xuB9cC/2HnWcBWhRP0VcA0Vz3gqcy/hDOkjYFksjni/BN4GXiXUmf83FXPQvUA/wjUZ2Q26sUjqzMweAN5193o/Q5DoMrOzgMnufmS6Y2msVEKXWjOzQ83sO7FT9BGEetN5aQ5LGrFYddb5wMx0x9KYKaHL7tiP0KTuG0Ib6vPc/fW0RiSNlpmdSLje8Bk1V+tINVTlIiISESqhi4hERNo65+rUqZPn5uama/ciIo3S4sWLv3D3zpUtS1tCz83NpbCwMF27FxFplMws8e7icqpyERGJCCV0EZGIUEIXEYmIjBqxaPv27RQVFbF169aaV5a0aN26NdnZ2bRs2TLdoYhIgoxK6EVFRey1117k5uZS9bgLki7uzrp16ygqKqJHjx7pDkdEEtRY5WJms8zsczNbUsVyM7ObzWxFbNiogbsbzNatW9lnn32UzDOUmbHPPvvoDEqSNns25OZCs2bheXaTHuK8/o9HMnXodwMjqll+EmHIqV6EYdFuq0tASuaZTd+PJGv2bJg8GVavBvfwPHly003qDXE8akzo7r6I0NVlVUYD93rwb8IoKV1TFaCI1F4mlIyvuAI2b644b/PmML+hNZXjkYpWLt2oOERXERWH0CpnZpPNrNDMCouLi1Ow69Rat24deXl55OXlsd9++9GtW7fy6W3btlX73sLCQi666KIa9zF06NAa1xGpi0wpGa9ZU7v59aVJHY9kBh4lDCW1pIpl/wSOjJteAAyqaZuDBg3yRMuWLdtlXnXuv989J8fdLDzff3+t3l6tqVOn+g033FBh3vbt21O3g0astt+TNKycHPeQuio+cnIURxTiAAq9HgeJLqLimIvZhLES61VD/epOnDiRSy+9lGOOOYbLL7+cV155haFDhzJgwACGDh3K8uXLAVi4cCEnn3wyANOmTWPSpEkMGzaMnj17cvPNN5dvr23btuXrDxs2jDFjxtCnTx8KCgrKfhCZP38+ffr04cgjj+Siiy4q3268VatWcdRRRzFw4EAGDhzIiy++WL7sd7/7Hf369aN///5MmTIFgBUrVjB8+HD69+/PwIED+eCDuowLLJksU0rG06dDVlbFeVlZYX5DalLHo6pMH/+g+hL6SMJ4hQYcDrySzDbrWkKv71/dshL6hAkTfOTIkV5SUuLu7uvXry8vqT/11FN+6qmnurv7s88+6yNHjix/75AhQ3zr1q1eXFzsHTt29G3btrm7e5s2bcrXb9euna9du9Z37Njhhx9+uD/33HO+ZcsWz87O9pUrV7q7+9ixY8u3G2/Tpk2+ZcsWd3d/7733vOx4zp8/34cMGeKbNm1yd/d169a5u/vgwYP94Ycfdnf3LVu2lC/fHSqhZ7ZMKZG61+9ZdLKidjyopoReYzt0Mysbo7GTmRURBp9tGfsxuJ0w0O0PCOMtbiaMT1jvGvJX9/TTT6d58+YArF+/ngkTJvD+++9jZmzfvr3S94wcOZJWrVrRqlUr9t13Xz777DOys7MrrDN48ODyeXl5eaxatYq2bdvSs2fP8nbe48aNY+bMXQdx2b59OxdccAFvvPEGzZs357333gPg6aef5uyzzyYrVhTo2LEjGzdu5KOPPuKUU04Bws1BEl3Tp4ez1fgLcOkoGQMUFIRHOjWl41FjQnf3cTUsd8JAvg2qe/dQzVLZ/FRr06ZN+eurrrqKY445hkceeYRVq1YxbNiwSt/TqlWr8tfNmzenpKQkqXU8yQFHbrzxRrp06cKbb75JaWlpeZJ2912aFia7TYmGsoRxxRWhgNO9e0he6U6s6dKUjkej7cslXfVz69evp1u30Ijn7rvvTvn2+/Tpw8qVK1m1ahUADzxQ+eD069evp2vXrjRr1oz77ruPHTt2AHDCCScwa9YsNseKI19++SXt2rUjOzubefPmAfDtt9+WL5doKiiAVaugtDQ8RzF51UZTOR6NNqEXFMDMmZCTA2bheebM+v+ifvWrX/HrX/+aI444ojyJptKee+7JrbfeyogRIzjyyCPp0qUL7du332W9888/n3vuuYfDDz+c9957r/wsYsSIEYwaNYr8/Hzy8vKYMWMGAPfddx8333wzhxxyCEOHDuXTTz9Neewikl5pG1M0Pz/fEwe4eOeddzjooIPSEk8m+eabb2jbti3uzs9//nN69erFJZdcku6wyul7EkkfM1vs7vmVLWu0JfQo+/Of/0xeXh4HH3ww69ev52c/+1m6QxKRRiCjeluU4JJLLsmoErmINA4qoYuIRIQSuohIRCihi4hEhBK6iEhEKKHHGTZsGE888USFeTfddBPnn39+te8pa375gx/8gK+//nqXdaZNm1beHrwq8+bNY9myZeXTV199NU8//XQtoheRpk4JPc64ceOYO3duhXlz585l3Lhqez8oN3/+fPbee+/d2ndiQr/22msZPnz4bm1LRJomJfQ4Y8aM4R//+AfffvstELqo/fjjjznyyCM577zzyM/P5+CDD2bq1KmVvj83N5cvvvgCgOnTp3PggQcyfPjw8i52IbQxP/TQQ+nfvz+nnXYamzdv5sUXX+TRRx/lsssuIy8vjw8++ICJEyfy0EMPAbBgwQIGDBhAv379mDRpUnl8ubm5TJ06lYEDB9KvXz/efffdXWJSN7siTUfGtkO/+GJ4443UbjMvD266qerl++yzD4MHD+bxxx9n9OjRzJ07lzPOOAMzY/r06XTs2JEdO3Zw3HHH8dZbb3HIIYdUup3Fixczd+5cXn/9dUpKShg4cCCDBg0C4NRTT+Wcc84B4Morr+TOO+/kwgsvZNSoUZx88smMGTOmwra2bt3KxIkTWbBgAb179+ass87itttu4+KLLwagU6dOvPbaa9x6663MmDGDO+64o8L79913X5566ilat27N+++/z7hx4ygsLOSxxx5j3rx5vPzyy2RlZfHll2GUwYKCAqZMmcIpp5zC1q1bKS0trf2BFpG0UAk9QXy1S3x1y4MPPsjAgQMZMGAAS5curVA9kui5557jlFNOISsri3bt2jFq1KjyZUuWLOGoo46iX79+zJ49m6VLl1Ybz/Lly+nRowe9e/cGYMKECSxatKh8+amnngrAoEGDyjv0ird9+3bOOecc+vXrx+mnn14ed7Ld7GYl9oCWwTJh3EiRdMrYEnp1Jen69KMf/YhLL72U1157jS1btjBw4EA+/PBDZsyYwauvvkqHDh2YOHEiW7durXY7iV3Ylpk4cSLz5s2jf//+3H333SxcuLDa7dTU105ZF7xVddHbVLrZLRvBqqwTybIRrCC6PeuJJFIJPUHbtm0ZNmwYkyZNKi+db9iwgTZt2tC+fXs+++wzHnvssWq3cfTRR/PII4+wZcsWNm7cyN///vfyZRs3bqRr165s376d2XFFyL322ouNGzfusq0+ffqwatUqVqxYAYReE7///e8n/XmaSje7mTTCvEi6KKFXYty4cbz55puMHTsWgP79+zNgwAAOPvhgJk2axBFHHFHt+wcOHMgZZ5xBXl4ep512GkcddVT5st/+9rccdthhHH/88fTp06d8/tixY7nhhhsYMGBAhQuRrVu35q677uL000+nX79+NGvWjHPPPTfpz9JUutnNlHEjRdJJ3edKrWXi95SbW/kIVjk5YUADkahQ97kSeZkywrxIOimhSySkawQrkUySca1cKmt9IZkjk1vCZMII8yLplFEl9NatW7Nu3bqMThpNmbuzbt268qaPIpJZMqqEnp2dTVFREcXFxekORarQunVrsrOz0x2GiFQioxJ6y5Yt6dGjR7rDEBFplDKqykVERHafErqISEQooYuIRIQSuohIRCihi4hEhBK6iEhEJJXQzWyEmS03sxVmNqWS5R3M7BEze8vMXjGz76U+VBERqU6NCd3MmgO3ACcBfYFxZtY3YbXfAG+4+yHAWcAfUx2oiIhUL5kS+mBghbuvdPdtwFxgdMI6fYEFAO7+LpBrZl1SGqmIiFQrmYTeDVgbN10UmxfvTeBUADMbDOQAu9wfbmaTzazQzAp1e7+ISGolk9Ar6/owsfes64EOZvYGcCHwOrDLAJfuPtPd8909v3PnzrWNVUREqpFMXy5FwAFx09nAx/EruPsG4GwAC33ffhh7iIhIA0mmhP4q0MvMepjZHsBY4NH4Fcxs79gygJ8Ci2JJXkREGkiNJXR3LzGzC4AngObALHdfambnxpbfDhwE3GtmO4BlwE/qMWYREalEUt3nuvt8YH7CvNvjXr8E9EptaCIiUhu6U1REJCKU0EVEIkIJXUQkIpTQRUQiQgldRCQilNBFUmz2bMjNhWbNwvPs2emOSJqKpJotikhyZs+GyZNh8+YwvXp1mAYoKEhfXNI0qIQukkJXXLEzmZfZvDnMF6lvSugiKbRmTe3mi6SSErpICnXvXrv5IqmkhC6SQtOnQ1ZWxXlZWWG+SH1TQpc6U6uOnQoKYOZMyMkBs/A8c6YuiErDUCsXqRO16thVQUHT/eySXiqhS52oVYdI5lBClzpRqw6RzKGELnWiVh0imUMJXepErTpEMocSutSJWnWIZA61cpE6U6sOkcygErqISEQooYuIRIQSuohIRCihi4hEhBK6iEhEKKGLiESEEnojpl4ORSSe2qE3UurlUEQSqYTeSKmXQxFJpITeSKmXQxFJpITeSKmXQxFJpITeSKmXQxFJlFRCN7MRZrbczFaY2ZRKlrc3s7+b2ZtmttTMzk59qBJPvRyKSCJz9+pXMGsOvAccDxQBrwLj3H1Z3Dq/Adq7++Vm1hlYDuzn7tuq2m5+fr4XFham4COIiDQdZrbY3fMrW5ZMCX0wsMLdV8YS9FxgdMI6DuxlZga0Bb4ESuoQs4iI1FIyCb0bsDZuuig2L96fgIOAj4G3gV+4e2nihsxsspkVmllhcXHxboYsIiKVSSahWyXzEutpTgTeAPYH8oA/mVm7Xd7kPtPd8909v3PnzrUMVUREqpNMQi8CDoibziaUxOOdDTzswQrgQ6BPakIUEZFkJJPQXwV6mVkPM9sDGAs8mrDOGuA4ADPrAhwIrExloCIiUr0a+3Jx9xIzuwB4AmgOzHL3pWZ2bmz57cBvgbvN7G1CFc3l7v5FPcYtIiIJkuqcy93nA/MT5t0e9/pj4ITUhiYiIrWhO0VFRCJCCV1EJCKU0EVEIkIJXUQkIpTQRUQiQgldRCQilNBFRCJCCV1EJCKU0EVEIkIJXUQkIpTQRUQiQgldRCQilNBFRCJCCV1EJCKU0EVEIkIJXUQkIpTQRUQiQgldRCQilNBFRCJCCV1EJCKU0EVEIkIJXUQkIpTQRUQiQgldRCQilNBFRCJCCV1EJCKU0EVEIkIJXUQkIpTQRUQiQgldRCQikkroZjbCzJab2Qozm1LJ8svM7I3YY4mZ7TCzjqkPNzPMng25udCsWXiePTvdEYmIJJHQzaw5cAtwEtAXGGdmfePXcfcb3D3P3fOAXwP/cvcv6yHetJs9GyZPhtWrwT08T56spC4i6ZdMCX0wsMLdV7r7NmAuMLqa9ccBf0lFcJnoiitg8+aK8zZvDvNFRNIpmYTeDVgbN10Um7cLM8sCRgB/q2L5ZDMrNLPC4uLi2saaEdasqd18EZGGkkxCt0rmeRXr/hB4oarqFnef6e757p7fuXPnZGPMKN27126+iEhDSSahFwEHxE1nAx9Xse5YIlzdAjB9OmRlVZyXlRXmi4ikUzIJ/VWgl5n1MLM9CEn70cSVzKw98H3g/1IbYmYpKICZMyEnB8zC88yZYb6ISDq1qGkFdy8xswuAJ4DmwCx3X2pm58aW3x5b9RTgSXffVG/RZoiCAiVwEck85l5VdXj9ys/P98LCwrTsW0SksTKzxe6eX9ky3SkqIhIRSugiIhGhhC4iEhFK6CIiEaGELiISEUroIiIRoYQuIhIRSugiIhGhhC4iEhFK6CIiEaGELiISEUroIiIRoYQuIhIRSugiIhGhhC4iEhFK6CIiEaGELiISEUroIiIRoYQuIhIRSugiIhGhhC4iEhFK6CIiEaGELiISEUroIiIRoYQuIhIRSugiIhGhhC4iEhFK6CIiEaGELiISEUroIiIRkVRCN7MRZrbczFaY2ZQq1hlmZm+Y2VIz+1dqwxQRkZq0qGkFM2sO3AIcDxQBr5rZo+6+LG6dvYFbgRHuvsbM9q2neEVEpArJlNAHAyvcfaW7bwPmAqMT1hkPPOzuawDc/fPUhikiIjVJJqF3A9bGTRfF5sXrDXQws4VmttjMzqpsQ2Y22cwKzaywuLh49yIWEZFKJZPQrZJ5njDdAhgEjAROBK4ys967vMl9prvnu3t+586dax2siEhjt2MHeGIGTZFkEnoRcEDcdDbwcSXrPO7um9z9C2AR0D81IYqIRMPatXD00TBrVv1sP5mE/irQy8x6mNkewFjg0YR1/g84ysxamFkWcBjwTmpDFRFpvObPhwED4K23oF27+tlHjQnd3UuAC4AnCEn6QXdfambnmtm5sXXeAR4H3gJeAe5w9yX1E7KISONRUgJTpsDIkZCdDa+9BqefXj/7Mq+vypwa5Ofne2FhYVr2LSLSED76CMaOheefh8mT4aabYM8967ZNM1vs7vmVLauxHbqINE7ffgtffw1duqQ7kqbpiSfgxz+GLVtg9mwYP77+96lb/0Uiwh3efRf++Mdwet+xI3TtCj/9KXyuO0MaTEkJXHEFjBgRjv/ixQ2TzEEldJFG7csvYcGCUBp88snQigKgd2+YNAmaNYNbb4WHHoJrroHzz4eWLdMbc5R9/HFI3v/6F/zkJ3DzzZCV1XD7V0IXaUS2b4eXXw7J+4kn4NVXQ8m8fXs47ji48ko4/njo0WPne847D37xC7j4Yvjzn0OSOfbYtH2EyHrqKSgogE2b4N574cwzGz4GJXSRDLdy5c4S+DPPwIYNoeR92GFw9dVwwgkweDC0qOK/uU8fePxxePRRuOSSkPjHjIEZMyAnp2E/SxTt2BHOfq67Dvr2hb/+FQ46KD2xqJWLSIbZsAGefXZnEv/ggzC/e3c48cTwOPZY6NCh9tvesgV+/3v4z/8M01OmwGWX1b3lRVP16aehiuXZZ2HiRPjTn6BNm/rdZ3WtXJTQRdJsx45w4aysGuWll8K8Nm3gmGNCCfzEE6FXL7DKOuLYDWvWhET+4IOQmwt/+AP86Eep235T8MwzIZlv2BCuU0yc2DD7VUIXyTBr14YE/uST8PTT4eImwMCBIXmfcAIMGQKtWtVvHM8+CxddBEuWwPDhoX49XdUFjcWOHaF65ZprQnXWX/8KBx/ccPtXO3SRNNu0CRYt2lkKfyfWMUbXrvDDH4YEfvzx0NB91h1zDLz+Otx2W6iPP+QQuPBCmDo1XGiVij77LLQtf/ppOOusUDKv7yqW2lAJXaQelJaGPjvKEvjzz8O2baHEffTRO0vh3/te5lRzFBeH9tN33BF+WK6/HiZMCBdgBRYuhHHjws1at9wCZ5+dnu+uuhK6viqRFPrmG/jv/w59dgwYAJdfHm7queCCkNi/+iok+f/4D+jXL3OSOYQkPnMmvPIK9OwZ2rEPHRqaRjZlpaWhiuW448JZyyuvhGOTSd9dGSV0kRTYvDk0A+zRI7Qc6dcP7roLiorg7bdDy5ITTmgcrUny8+GFF+Cee2DVqtAk8ic/CdUNTc3nn4c7Pq+6KpTOCwvDd5uplNBF6mDLFrjxxpDIL7sslMpffDGUxidOhG6JY3s1Es2ahTri996DX/4y3CjTu3foXGr79nRH1zAWLQrf56JF4czlvvugbdt0R1U9JXSR3bB1a2gR0rMnXHppqAt/7rlQnTJkSLqjS5127eCGG8JZxuGHhxuT8vJCdwNRVVoK//Vf4YJxmzbhztxzzsnMKpZESugitfDtt+GC2He+E26nP/DA0G/HggVw5JHpjq7+lN1tOm9eOCsZPjzcbbpqVbojS60vvggdm/3mN/D//l+4P6B/Ixp7TQldJAnbtsHtt8N3vxsucPbsGW4sWbgwtFppCsxg9GhYtgx++9swAs9BB4X22Fu2pDu6unvhhXD28eyzoRnnnDmw117pjqp2Gl1CX7o0lIQWLkx3JNIUbN8eOrTq1St0cnXAAaETpkWLwil5U9S6degE7N13YdQomDYtJPaHH66/wY/rU2kp/O538P3vh8/20ktw7rmNo4olUaNL6J98AqtXh3+mESPCcE4iqbZ9exjIt3fvMNJM166hyuGFF0J1Q2P8Z0+17t3hgQdCiXavveC000JLnmXL0h1Z8tatCz9Kl18Op54a8smAAemOavc1uoQ+fHi48j5jRmhCNGhQqOtavjzdkUkUlJSE5np9+oSmep06wT//GUptJ56oRF6ZYcPC3aY33xz+Jw85JFw8Xb8+3ZFV76WXQvJ+6qnQqdYDD9Tf4M0Nxt3T8hg0aJDX1fr17ldf7d6mjXvz5u4//an72rV13qw0QSUl7vfd596rlzu4Dxjg/uij7qWl6Y6scfn8c/dzznE3c993X/dZs9x37Eh3VBWVlrrPmOHeooV7jx7uhYXpjqh2gEKvIq9G4tb/zz8P3YHedlsoQV1wAfz617DPPinZvETYjh2hZHbtteEsr3//UCc8erRK43WxeHHoE+all8KNSZdfDnvvHbo+aNUq1FWXvU581GdXA19+Ge4P+PvfQxXLnXeGuBqTJtPb4urV4Z/x3nvDDQC//GU49cv0mwGk4ZWWhl7yrrkmdJT1ve+Fv51TTlHfJalSWhoGR/7Vr0K/4clq2bLqZF/Tj0F1ywCmTw/DxM2YEX5wGuOPdpNJ6GWWLQtX4R95JPRPceWV8LOf1X9XpA3JPXT+NHcufPRRuECVk7Pz0b1747jNvKGVlobWGNdcE7qM7ds3JPLTTlMiry/ffBOO9datoR1/ZY/dXVbV8m3bKo8lNzeckQ0e3KCHIKWaXEIv8/LL4QaBZ54JSe6aa0LXl82b1+tu69UHH8Bf/hIey5aFz9K1ayh1lJZWXLdz54pJPjHhd+jQOEsou8M93BQzbVr4ITzwwPD69NMb99+DVM49JPXEhJ+dHUrxjVmTTehlnn461KkXFoYS2XXXNa7RWT79NIwsM2dO+JECOOqoMFrKmDGhJUZJSSipr16987FmTcXprVsrbrdt2+oTfteujb/U6h7qS6dNCy0xevUKfX2PHatELo1Tk0/oEP6xH3545w0RgweH/hoydfTzr78OVUZz5oQzjNLScBfb+PFwxhkh4daGe+jvOj7BJyb9r76q+J6WLcONNFUl/AMOyNxqLPdwJ+O0aeGH/DvfCQM4jB9f9WDKIo1BZBL67NmhA/41a0JCmT4dCgpqt9+SknDRdNq0MAzY8ceHFjL5lR6ehrVlS2jzPGdOeN62LSSi8eND1531PTTYxo27Jvz4pP/JJxXvBDSD/fbbmeQ7dIA99giPVq12vq7rdNnrli1rPmNwDz0dTp0a+q3u0SN0fXrmmUrkEg2RSOizZ4c79jZv3jkvKyt0a1nbpA6h+uH228OPwhdfhIti110XbihpSCUloWOnOXNCiXzjxpAkx44NSfzQQzOnaujbb0P/3lUl/A0bwo9Q2WPHjtTH0KJF9cl/y5bQ/DAnJ5yNTZgQfghEoiISCT03NySNRDk5devxbcOG0J/1jBnhx2LixFC6q22VRm24h/a5c+aEuvHi4jASymmnhdL4sGHRqN/dsaNigi9rfVDVdDLr1DRdUhJu5T777JDgRaImEgm9WbPKO/4x27V1x+4oLg516rfcErZ5/vmhhUynTnXfdpm33w5JfO7c8CPUunUYIHj8eDjppMytjxaRzFHnMUXNbISZLTezFWY2pZLlw8xsvZm9EXtcXdegE1VVYk5VSbpzZ/jDH+D990MVzh//GLpIveaaUA2yuz78MNTR9+sX+ri44YZQF37vveEO1wcfDC1ulMxFpK5qTOhm1hy4BTgJ6AuMM7O+laz6nLvnxR7XpjhOpk8PdebxsrLC/FTq3j3cDrxkSeg5btq0kNhvumnXZn9V+ewz+J//CSPX9OwZLuS2bx9K/598ElpfnHlm4+trWUQyWzIl9MHACndf6e7bgLnA6PoNa1cFBeECaE5OqBLJydn9C6LJOOggeOih0FIiLy90IXDggWHg35KSXddfvx7uvjv8COy/P1x0UaiTv/76UL3y/POhGqdz5/qJV0QkmYTeDVgbN10Um5doiJm9aWaPmdnBlW3IzCabWaGZFRYXF9c62IKCkBxLS8NzfSXzeIceGrrXXLAgtD6ZNClUnzz8cGhR8be/hZt7unQJF+JWrAg3MS1ZAm++GTolysmp/zhFRJJpmVtZo7nEy5OvATnu/o2Z/QCYB/Ta5U3uM4GZEC6K1i7U9Dr2WPj3v8Pt41dcEVqktGwZBkLo0iX0FTN+fLhhKVOaGYpI05JMQi8CDoibzgY+jl/B3TfEvZ5vZreaWSd3/yI1YWYGs9Ab36hRcP/9oYvQUaNCM0PdtCIi6ZZMGnoV6GVmPYCPgLHA+PgVzGw/4DN3dzMbTKjKWZfqYDNF8+bhhpUJE9IdiYjITjUmdHcvMbMLgCeA5sAsd19qZufGlt8OjAHOM7MSYAsw1tPVwF1EpIlqNDcWiYhICm4sEhGRzKeELiISEUroIiIRoYQuIhIRSugiIhGhhC4iEhFpa7ZoZsVAJUNWNCqdgEjdDVtHOh4V6XjspGNRUV2OR467V9rNX9oSehSYWWFV7UGbIh2PinQ8dtKxqKi+joeqXEREIkIJXUQkIpTQ62ZmugPIMDoeFel47KRjUVG9HA/VoYuIRIRK6CIiEaGELiISEUrou8HMDjCzZ83sHTNbama/SHdM6WZmzc3sdTP7R7pjSTcz29vMHjKzd2N/I0PSHVM6mdklsf+TJWb2FzNrne6YGpKZzTKzz81sSdy8jmb2lJm9H3vukIp9KaHvnhLgP9z9IOBw4Odm1jfNMaXbL4B30h1Ehvgj8Li79wH604SPi5l1Ay4C8t39e4RBcsamN6oGdzcwImHeFGCBu/cCFsSm60wJfTe4+yfu/lrs9UbCP2y39EaVPmaWDYwE7kh3LOlmZu2Ao4E7Adx9m7t/ndag0q8FsKeZtQCySBiTOOrcfRHwZcLs0cA9sdf3AD9Kxb6U0OvIzHKBAcDLaQ4lnW4CfgWUpjmOTNATKAbuilVB3WFmbdIdVLq4+0fADGAN8Amw3t2fTG9UGaGLu38CoYAI7JuKjSqh14GZtQX+Blzs7hvSHU86mNnJwOfuvjjdsWSIFsBA4DZ3HwBsIkWn041RrG54NNAD2B9oY2Y/Tm9U0aWEvpvMrCUhmc9294fTHU8aHQGMMrNVwFzgWDO7P70hpVURUOTuZWdsDxESfFM1HPjQ3YvdfTvwMDA0zTFlgs/MrCtA7PnzVGxUCX03mJkR6kjfcfc/pDuedHL3X7t7trvnEi52PePuTbYE5u6fAmvN7MDYrOOAZWkMKd3WAIebWVbs/+Y4mvBF4jiPAhNirycA/5eKjbZIxUaaoCOAM4G3zeyN2LzfuPv89IUkGeRCYLaZ7QGsBM5Oczxp4+4vm9lDwGuE1mGv08S6ATCzvwDDgE5mVgRMBa4HHjSznxB+9E5Pyb5067+ISDSoykVEJCKU0EVEIkIJXUQkIpTQRUQiQgldRCQilNBFRCJCCV1EJCL+P1Qg+K408FL8AAAAAElFTkSuQmCC\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",
    "acc = history.history['acc']\n",
    "val_acc = history.history['val_acc']\n",
    "loss = history.history['loss']\n",
    "val_loss = history.history['val_loss']\n",
    "epochs = range(1, len(acc) + 1)\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",
    "plt.figure()\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",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "89b86631",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "418178b4",
   "metadata": {},
   "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": "3.8.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}