model_training.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From c:\\Users\\pcd\\Documents\\GitHub\\elden-ring-dialogue-generator\\.venv\\lib\\site-packages\\keras\\src\\losses.py:2976: The name tf.losses.sparse_softmax_cross_entropy is deprecated. Please use tf.compat.v1.losses.sparse_softmax_cross_entropy instead.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "import csv\n",
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras.utils import to_categorical\n",
    "from tensorflow.keras.preprocessing.text import Tokenizer\n",
    "from tensorflow.keras.preprocessing.sequence import pad_sequences\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "There are 8014 lines of dialogue\n",
      "\n",
      "The first 5 lines look like this:\n",
      "\n",
      "text\n",
      "The Tarnished will soon return.\n",
      "Guided by Grace once lost.\n",
      "The Golden Order is broken to its core.\n",
      "They will fight. And they will die. In an unending curse.\n"
     ]
    }
   ],
   "source": [
    "\n",
    "\n",
    "corpus = []\n",
    "with open('ER-Data.csv', 'r') as csvfile:\n",
    "    reader = csv.reader(csvfile)\n",
    "    for row in reader:\n",
    "        corpus.append(row[1])\n",
    "    \n",
    "print(f\"There are {len(corpus)} lines of dialogue\\n\")\n",
    "print(f\"The first 5 lines look like this:\\n\")\n",
    "for i in range(5):\n",
    "  print(corpus[i])    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Fit the tokenizer on the data\n",
    "tokenizer = Tokenizer()\n",
    "tokenizer.fit_on_texts(corpus)\n",
    "total_words = len(tokenizer.word_index) + 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "There are 54010 n-grams\n",
      "\n",
      "The longest n-gram has 30 words\n",
      "\n",
      "The first 10 n-grams look like this:\n",
      "\n",
      "[1, 40]\n",
      "[1, 40, 22]\n",
      "[1, 40, 22, 238]\n",
      "[1, 40, 22, 238, 239]\n",
      "[775, 36]\n",
      "[775, 36, 114]\n",
      "[775, 36, 114, 83]\n",
      "[775, 36, 114, 83, 328]\n",
      "[1, 131]\n",
      "[1, 131, 121]\n"
     ]
    }
   ],
   "source": [
    "# Generate n-grams on the data\n",
    "input_sequences = []\n",
    "for line in corpus:\n",
    "    sequence = tokenizer.texts_to_sequences([line])[0]\n",
    "    for i in range(1, len(sequence)):\n",
    "        n_gram_sequence = sequence[:i+1]\n",
    "        input_sequences.append(n_gram_sequence)\n",
    "\n",
    "# Save max sequence length\n",
    "max_sequence_len = max([len(x) for x in input_sequences])\n",
    "        \n",
    "print(f\"There are {len(input_sequences)} n-grams\\n\")\n",
    "print(f\"The longest n-gram has {max_sequence_len} words\\n\")\n",
    "print(f\"The first 10 n-grams look like this:\\n\")\n",
    "for i in range(10):\n",
    "  print(input_sequences[i])        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Padded corpus has shape (54010, 30)\n",
      "\n",
      "The first 10 padded n-grams look like this:\n",
      "\n",
      "[ 0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0\n",
      "  0  0  0  0  1 40]\n",
      "[ 0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0\n",
      "  0  0  0  1 40 22]\n",
      "[  0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0\n",
      "   0   0   0   0   0   0   0   0   1  40  22 238]\n",
      "[  0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0\n",
      "   0   0   0   0   0   0   0   1  40  22 238 239]\n",
      "[  0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0\n",
      "   0   0   0   0   0   0   0   0   0   0 775  36]\n",
      "[  0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0\n",
      "   0   0   0   0   0   0   0   0   0 775  36 114]\n",
      "[  0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0\n",
      "   0   0   0   0   0   0   0   0 775  36 114  83]\n",
      "[  0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0\n",
      "   0   0   0   0   0   0   0 775  36 114  83 328]\n",
      "[  0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0\n",
      "   0   0   0   0   0   0   0   0   0   0   1 131]\n",
      "[  0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0\n",
      "   0   0   0   0   0   0   0   0   0   1 131 121]\n"
     ]
    }
   ],
   "source": [
    "# pad sequences\n",
    "input_sequences = np.array(pad_sequences(input_sequences, maxlen=max_sequence_len, padding='pre'))\n",
    "print(f\"Padded corpus has shape {input_sequences.shape}\\n\")\n",
    "print(f\"The first 10 padded n-grams look like this:\\n\")\n",
    "for i in range(10):\n",
    "  print(input_sequences[i])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Feutures have shape (54010, 29)\n",
      "\n",
      "Labels have shape (54010, 5195)\n",
      "\n",
      "The first 5 feutures look like this:\n",
      "\n",
      "[0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1]\n",
      "[ 0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0\n",
      "  0  0  0  1 40]\n",
      "[ 0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0\n",
      "  0  0  1 40 22]\n",
      "[  0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0\n",
      "   0   0   0   0   0   0   0   1  40  22 238]\n",
      "[  0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0   0\n",
      "   0   0   0   0   0   0   0   0   0   0 775]\n",
      "The first 5 labels look like this:\n",
      "\n",
      "[0. 0. 0. ... 0. 0. 0.]\n",
      "[0. 0. 0. ... 0. 0. 0.]\n",
      "[0. 0. 0. ... 0. 0. 0.]\n",
      "[0. 0. 0. ... 0. 0. 0.]\n",
      "[0. 0. 0. ... 0. 0. 0.]\n"
     ]
    }
   ],
   "source": [
    "# Split the n-grams into feutures and labels\n",
    "# The last word in each n-gram is the label\n",
    "feutures, labels = input_sequences[:,:-1],input_sequences[:,-1]\n",
    "\n",
    "# One-hot encode the labels\n",
    "labels = to_categorical(labels, num_classes=total_words)\n",
    "\n",
    "print(f\"Feutures have shape {feutures.shape}\\n\")\n",
    "print(f\"Labels have shape {labels.shape}\\n\")\n",
    "print(f\"The first 5 feutures look like this:\\n\")\n",
    "for i in range(5):\n",
    "  print(feutures[i])\n",
    "print(f\"The first 5 labels look like this:\\n\")\n",
    "for i in range(5):\n",
    "  print(labels[i])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From c:\\Users\\pcd\\Documents\\GitHub\\elden-ring-dialogue-generator\\.venv\\lib\\site-packages\\keras\\src\\backend.py:873: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.\n",
      "\n",
      "WARNING:tensorflow:From c:\\Users\\pcd\\Documents\\GitHub\\elden-ring-dialogue-generator\\.venv\\lib\\site-packages\\keras\\src\\optimizers\\__init__.py:309: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Build the model\n",
    "model = tf.keras.Sequential([\n",
    "    tf.keras.layers.Embedding(total_words, 100, input_length=max_sequence_len-1),\n",
    "    tf.keras.layers.Bidirectional(tf.keras.layers.LSTM(64, activation='relu', return_sequences=True)),\n",
    "    tf.keras.layers.Bidirectional(tf.keras.layers.LSTM(64, activation='relu')),\n",
    "    tf.keras.layers.Dense(total_words, activation='softmax')\n",
    "])\n",
    "\n",
    "model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/50\n",
      "WARNING:tensorflow:From c:\\Users\\pcd\\Documents\\GitHub\\elden-ring-dialogue-generator\\.venv\\lib\\site-packages\\keras\\src\\utils\\tf_utils.py:492: The name tf.ragged.RaggedTensorValue is deprecated. Please use tf.compat.v1.ragged.RaggedTensorValue instead.\n",
      "\n",
      "WARNING:tensorflow:From c:\\Users\\pcd\\Documents\\GitHub\\elden-ring-dialogue-generator\\.venv\\lib\\site-packages\\keras\\src\\engine\\base_layer_utils.py:384: The name tf.executing_eagerly_outside_functions is deprecated. Please use tf.compat.v1.executing_eagerly_outside_functions instead.\n",
      "\n",
      "1688/1688 [==============================] - 35s 19ms/step - loss: 6.5192 - accuracy: 0.0588\n",
      "Epoch 2/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 5.9045 - accuracy: 0.0894\n",
      "Epoch 3/50\n",
      "1688/1688 [==============================] - 31s 18ms/step - loss: 5.5228 - accuracy: 0.1185\n",
      "Epoch 4/50\n",
      "1688/1688 [==============================] - 31s 18ms/step - loss: 5.2055 - accuracy: 0.1424\n",
      "Epoch 5/50\n",
      "1688/1688 [==============================] - 36s 21ms/step - loss: 4.9106 - accuracy: 0.1612\n",
      "Epoch 6/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 4.6245 - accuracy: 0.1815\n",
      "Epoch 7/50\n",
      "1688/1688 [==============================] - 34s 20ms/step - loss: 4.3365 - accuracy: 0.2016\n",
      "Epoch 8/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 4.0469 - accuracy: 0.2300\n",
      "Epoch 9/50\n",
      "1688/1688 [==============================] - 35s 21ms/step - loss: 3.7510 - accuracy: 0.2590\n",
      "Epoch 10/50\n",
      "1688/1688 [==============================] - 37s 22ms/step - loss: 3.4677 - accuracy: 0.2924\n",
      "Epoch 11/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 3.1971 - accuracy: 0.3277\n",
      "Epoch 12/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 2.9550 - accuracy: 0.3648\n",
      "Epoch 13/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 2.7410 - accuracy: 0.3986\n",
      "Epoch 14/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 2.5557 - accuracy: 0.4305\n",
      "Epoch 15/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 2.3949 - accuracy: 0.4573\n",
      "Epoch 16/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 2.2458 - accuracy: 0.4854\n",
      "Epoch 17/50\n",
      "1688/1688 [==============================] - 31s 19ms/step - loss: 2.1215 - accuracy: 0.5083\n",
      "Epoch 18/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 2.0030 - accuracy: 0.5300\n",
      "Epoch 19/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 1.8984 - accuracy: 0.5513\n",
      "Epoch 20/50\n",
      "1688/1688 [==============================] - 31s 19ms/step - loss: 1.8099 - accuracy: 0.5681\n",
      "Epoch 21/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 1.7207 - accuracy: 0.5871\n",
      "Epoch 22/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 1.6448 - accuracy: 0.6019\n",
      "Epoch 23/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 1.5677 - accuracy: 0.6168\n",
      "Epoch 24/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 1.5116 - accuracy: 0.6300\n",
      "Epoch 25/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 1.4557 - accuracy: 0.6405\n",
      "Epoch 26/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 1.3967 - accuracy: 0.6530\n",
      "Epoch 27/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 1.3584 - accuracy: 0.6615\n",
      "Epoch 28/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 1.2995 - accuracy: 0.6747\n",
      "Epoch 29/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 1.2605 - accuracy: 0.6834\n",
      "Epoch 30/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 1.2282 - accuracy: 0.6907\n",
      "Epoch 31/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 1.2037 - accuracy: 0.6972\n",
      "Epoch 32/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 1.1621 - accuracy: 0.7049\n",
      "Epoch 33/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 1.1363 - accuracy: 0.7099\n",
      "Epoch 34/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 1.1143 - accuracy: 0.7146\n",
      "Epoch 35/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 1.0891 - accuracy: 0.7225\n",
      "Epoch 36/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 1.0686 - accuracy: 0.7256\n",
      "Epoch 37/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 1.0686 - accuracy: 0.7264\n",
      "Epoch 38/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 1.0385 - accuracy: 0.7328\n",
      "Epoch 39/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 0.9919 - accuracy: 0.7453\n",
      "Epoch 40/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 0.9854 - accuracy: 0.7451\n",
      "Epoch 41/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 0.9878 - accuracy: 0.7448\n",
      "Epoch 42/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 0.9575 - accuracy: 0.7506\n",
      "Epoch 43/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 0.9538 - accuracy: 0.7508\n",
      "Epoch 44/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 0.9292 - accuracy: 0.7591\n",
      "Epoch 45/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 0.9279 - accuracy: 0.7569\n",
      "Epoch 46/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 0.9137 - accuracy: 0.7624\n",
      "Epoch 47/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 0.9074 - accuracy: 0.7636\n",
      "Epoch 48/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 1.0358 - accuracy: 0.7318\n",
      "Epoch 49/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 0.8745 - accuracy: 0.7717\n",
      "Epoch 50/50\n",
      "1688/1688 [==============================] - 32s 19ms/step - loss: 0.8471 - accuracy: 0.7816\n"
     ]
    }
   ],
   "source": [
    "history = model.fit(feutures, labels, epochs=50, verbose=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAiMAAAGdCAYAAADAAnMpAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjguNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8fJSN1AAAACXBIWXMAAA9hAAAPYQGoP6dpAAA70klEQVR4nO3de1xVVf7/8TeggKaAhoISZpcxx0owUCKnqxRTZtqVrhpTNpVaSv1KMzHtgt2MUhLH7J5JOtpUGl0obUxKw8zG0u7iDZBKIFIQzv79sb5CJCgHOexzeT0fj/1wnX32PnzYo533rL32Wn6WZVkCAACwib/dBQAAAN9GGAEAALYijAAAAFsRRgAAgK0IIwAAwFaEEQAAYCvCCAAAsBVhBAAA2Kqd3QU0h8Ph0I4dO9S5c2f5+fnZXQ4AAGgGy7JUUVGhnj17yt+/6f4PjwgjO3bsUHR0tN1lAACAFti6dauOOuqoJt/3iDDSuXNnSeaXCQkJsbkaAADQHOXl5YqOjq77Hm+KR4SR/bdmQkJCCCMAAHiYQw2xYAArAACwVYvCSFZWlnr37q3g4GAlJCRozZo1Bz0+MzNTJ5xwgjp06KDo6GhNmDBBe/fubVHBAADAuzgdRnJycpSWlqapU6dq3bp1iomJUXJyskpKSho9fsGCBZo4caKmTp2qr7/+WvPnz1dOTo7uueeewy4eAAB4PqfDyMyZMzV69GilpqaqX79+ys7OVseOHfXss882evzq1as1ePBgXX311erdu7fOO+88XXXVVYfsTQEAAL7BqTBSXV2tgoICJSUl1X+Av7+SkpKUn5/f6DmnnXaaCgoK6sLHDz/8oOXLl+uCCy5o8udUVVWpvLy8wQYAALyTU0/TlJaWqra2VhEREQ32R0REaNOmTY2ec/XVV6u0tFR/+9vfZFmWampqdPPNNx/0Nk1GRoamTZvmTGkAAMBDufxpmhUrVuihhx7S008/rXXr1mnJkiVatmyZ7r///ibPmTRpksrKyuq2rVu3urpMAABgE6d6RsLDwxUQEKDi4uIG+4uLixUZGdnoOVOmTNF1112nG2+8UZJ08sknq7KyUjfddJMmT57c6PSwQUFBCgoKcqY0AADgoZzqGQkMDFRcXJzy8vLq9jkcDuXl5SkxMbHRc37//fcDAkdAQIAkM2c9AADwbU7PwJqWlqZRo0YpPj5egwYNUmZmpiorK5WamipJGjlypKKiopSRkSFJGjZsmGbOnKkBAwYoISFB3333naZMmaJhw4bVhRIAAOC7nA4jKSkp2rVrl9LT01VUVKTY2Fjl5ubWDWotLCxs0BNy7733ys/PT/fee6+2b9+ubt26adiwYXrwwQdb77cAAAAey8/ygHsl5eXlCg0NVVlZGWvTAADgIZr7/c3aNAAA+KjaWmnBAmn4cNO2C2EEAAAfU1srvfqqdNJJ0jXXSG+8IeXk2FeP02NGAACAZ6qtlRYtkqZPl77+2uzr0kW64w7pwgvtq4swAgCAl3M46kPIV1+ZfWFhJoTcdptk93BMwggAAF7K4ZAWLzYhZONGsy8sTEpLMyEkNNTW8uoQRgAA8ELV1VJSkvTf/5rXoaH1ISQszNbSDkAYAQDAC734ogkiRxwh3XmnNH68+4WQ/QgjAAB4mX37pP1zi06fbnpE3BlhBAAAN1BTI5WXS2VlUlWV1KeP1Mhass3y0kvSTz9J3btLN9/cqmW6BGEEAAAXsyxp0yYpN1f67DPp119N6Ni92/xZVib99lvDc8aPl554wvmf9cdekbvukjp2PNzqXY8wAgDAIdTUmC04uPnnlJVJH3xgAkhurlRY2LzzOnSQ9uyRnnxSuuoqadAg52p95RXphx+kbt08o1dEIowAANCkb76Rnn5aev55Ey7Cw6WjjpKioxv+ub9dXl4fPlavbjjFemCgdOaZ0tlnS5GR5umWsDDz5/52SIg57rrrpJdflv75T2ntWqldM7+ta2qkBx4w7f/3/8zgVU/AQnkAAPxBTY301lsmhLz33uF9Vp8+UnKy9Pe/myDS3HBQUiL17Wtu5zz+ePMHoL7wgnT99SY0/fST/WGkud/f9IwAACATAJ55RsrOlrZuNfv8/KQLLpDGjDG3S7ZvN+9t22a2/e2tW83m7y8NGWLCR3KydOyxLaule3fpkUek0aOl9HTpssukXr0Ofs4fe0XuvNP+IOIMekYAAF7JsqQffzTjLxwOc8uksT8rKswtkUWLzOBPSTrySOmGG8yYi2OOaf7Pk0yAaQ0Oh3TGGdLHH5tVdV9//eDHv/SSNHKkqf2nn6ROnVqnjsNBzwgAwOdUVkp5edKyZWbbvt258wcNkm69VUpJcW6wqtR6IWQ/f39p7lwpNlb6z39MGBkxovFja2vre0XuuMM9gogzCCMAAI/2ww/14WPFCjNHx36BgWZQaECA+XL3969v//HPU081t2Li4237NRp14olmIGpGhjRunLkF1LnzgcctXGgG23btKo0d2/Z1Hi7CCACgTfzyi/TQQ9LmzQeGgT8HhHbt6rf27Rtvl5ZKy5eb+Tv+6JhjpKFDzXbWWc73cLibe+81YePHH6WpU6WZMxu+X1sr3X+/aaelNR5W3B1jRgAALmVZ0muvmQXaSkpa//MDAqS//U268EITQPr2bf1bJnbLzZXOP98EtbVrpVNOqX/v1Velq6+WunQxY0Xc6WuSMSMAANsVFpoxGMuWmdd//asJJQEBjQ8o3d/eP8nY/m3fvgPbgYFmzo7zznPfBeBay9//bsax5OSYuUc++aT+Gk6fbo6ZMMG9gogzCCMAgFZXWyvNni1NnmwGlQYGSvfcI02cKAUF2V2dZ3riCentt8108nPmmLEhixaZ21RhYSbkeaoWLsEDAEDjNmyQTjvNrK1SWWluoaxfb8Y7EERarkcPacYM077nHjOvyf6xIuPHm1lcPRVhBADQKvbskSZNkuLipDVrzC2D7Gxp5UpzewaH75//lBISzNwoZ50lffWVCSG33253ZYeH2zQAgBazLNMTsmSJ9OKLZgClJF1yiTRrltSzp63leZ39c4/ExZlHmiXTK+LpY2YIIwAApzgcpudjyRKzff99/Xs9e0pZWU1PzoXDFxNjBqs+9pjpffL0XhGJMAIAaIaaGmnVKunf/5aWLm04s2lwsFmH5ZJLpIsv9sx5LjzNtGmmV+qss8wjvZ6OMAIAPsaypOJi8xTG5s3m1kplpfT772ZrrF1SIpWV1X9Gp05mXo9LLjHzX3ja9OOermNH0zPiLQgjAOClamqkb781oWN/8Njf/mOwaK6uXc2CbZdcIiUlef7MpnAfhBEA8DJbtkjz5knPPivt3Nn4Mf7+Ztr0vn3NMvchIeb/bR9xhPnzz+3OnaV+/cx07EBrI4wAgBeoqTGznM6da6YO37/QR6dOJnD8eTv+eOb8gPsgjACAB9u6VXrmGWn+/IaDSocMMXNSDB9uZj8F3BlhBAA8TEmJ9NFH0gsvmFVrHQ6zv1s3KTVVGj3a9HwAnoIwAgBuzLLMwNOPPzaP1n78sRmU+kdnn216QUaM4NYLPBNhBADcyL59ZiG0/cHj44+l0tKGx/j5SSedZOb2GD1a6tPHnlqB1kIYAQAb1dRIn38uffih2f77XzO3xx8FB5v1SAYPNovOJSZ6/vTfwB+1KIxkZWXp0UcfVVFRkWJiYjRr1iwNGjSo0WPPOussrVy58oD9F1xwgZYtW9aSHw8AHsvhkL74oj58fPSRVF7e8Jgjj5ROP90Ej8GDpVNOYRAqvJvTYSQnJ0dpaWnKzs5WQkKCMjMzlZycrM2bN6t79+4HHL9kyRJVV1fXvf75558VExOjyy+//PAqBwAP8vnn5qmXnBzp558bvhcaKp15phn7cfbZ0sknm3lAAF/hZ1n7n0ZvnoSEBA0cOFCzZ8+WJDkcDkVHR2vcuHGaOHHiIc/PzMxUenq6du7cqSOOOKJZP7O8vFyhoaEqKytTSEiIM+UCgG3KyqQFC0wIWbeufn+nTqbn45xzTPiIjZUCAmwrE3CZ5n5/O9UzUl1drYKCAk2aNKlun7+/v5KSkpSfn9+sz5g/f76uvPLKgwaRqqoqVVVV1b0u/3MfJgC4Kcsyg07nzZMWLZL27DH7AwPNInI33GAWN2MmU6CeU2GktLRUtbW1ioiIaLA/IiJCmzZtOuT5a9as0f/+9z/Nnz//oMdlZGRo2rRpzpQGALbasaO+F2Tz5vr9/fqZJ16uvVYKD7evPsCdtenTNPPnz9fJJ5/c5GDX/SZNmqS0tLS61+Xl5YqOjnZ1eQDglMJC6d//NtvHH9fv79hRuvJK6cYbpVNPNY/iAmiaU2EkPDxcAQEBKi4ubrC/uLhYkZGRBz23srJSCxcu1PTp0w/5c4KCghTEzD0A3NAPP5jwsXixtGZNw/cSE80MqCkpZuE5AM3jVBgJDAxUXFyc8vLyNGLECElmAGteXp7Gjh170HMXLVqkqqoqXXvttS0uFgDs8P330muvmQDyx4Gofn5mIOqll0qXXCIddZR9NQKezOnbNGlpaRo1apTi4+M1aNAgZWZmqrKyUqmpqZKkkSNHKioqShkZGQ3Omz9/vkaMGKEjjzyydSoHABf69VczAPWFF6TVq+v3+/ubAaiXXWYGpB6iUxhAMzgdRlJSUrRr1y6lp6erqKhIsbGxys3NrRvUWlhYKP8/PSC/efNmrVq1Su+++27rVA0ALrBvn/TOO9KLL0pvvCHtf6jP39+sgnvFFWYV3G7d7K0T8DZOzzNiB+YZAeAqlmUmJHvxRfM0zK5d9e+ddJI0apR09dVSz5721Qh4KpfMMwIA3mLvXumVV6SnnpI2bKjf3727dM010siRUkwMT8IAbYEwAsCnFBVJTz8tZWfX94IEBUkjRpgAct55Ujv+ywi0Kf7JAfAJX3whPfGE9Oqr0v7lsnr1ksaNM7Oidulib32ALyOMAPBaDof01ltSZqZZIXe/U0+VJkwwj+PSCwLYj3+GALxSXp50663SN9+Y1wEB5nHc8eNNGAHgPggjALxKRYV0111mTIgkhYVJN90kjRljbssAcD+EEQBeIy/PjP/YssW8vvVWacYMqXNne+sCcHCEEQAe78+9Ib17S/PnS+ecY2tZAJrJ/9CHAID7+uAD6eST64PIrbdKX35JEAE8CT0jADwSvSGA96BnBIBH2btXev55qX9/ekMAb0HPCACPsH27NGeONHeuVFpq9tEbAngHwggAt2VZ0iefmPVjFi+WamrM/l69zKO6t94qdepkb40ADh9hBIDbqaqSFi2SnnxS+uyz+v1nnCHddps0fDgzpwLehH/OANxGba1ZxO7BB6XiYrMvKEi6+mqzhsyAAfbWB8A1CCMA3ML69Wam1LVrzeuePc1tmJtukrp1s7U0AC5GGAFgq8pK6b77zIq6tbVSaKiUkSHdeKPUvr3d1QFoC4QRALbJzZVuuUX66Sfz+vLLzTiRHj1sLQtAG2OeEQBtrrhYuuoq6fzzTRCJjpbefFN67TWCCOCLCCMA2ozDIT3zjNS3r7RwoeTvL02YIH31lXThhXZXB8Au3KYB0Cb++19p4kRp9WrzesAAad48KS7O3roA2I+eEQAu9cUX0tChZo6Q1auljh2lxx6T1qwhiAAw6BkB4BLffy+lp0sLFpjXAQHmCZkpU6SoKHtrA+BeCCMAWtXOndL995tbMPunb7/ySmn6dOkvf7G3NgDuiTACoFXs3i098oiUmSnt2WP2/f3v0kMPMXMqgIMjjAA4LJYl5eRIY8dKP/9s9iUmmonLzjzT3toAeAbCCIAWKy42U7YvWWJe9+tnQsiwYZKfn721AfAchBEATrMsM0HZmDGmN6RdO+nee6V77mEKdwDOI4wAcEpJiZnCfX9vSGys9PzzUkyMnVUB8GTMMwKg2V57zdyKWbLE9IZMnSp9+ilBBMDhoWcEwCGVlJhbMosXm9f9+0svvGB6RQDgcNEzAuCgli+XTjzRBJF27cxEZmvXEkQAtB56RgA0qqbGDEp9+GHz+uSTTW8Ic4YAaG2EEQAH2LZNuuoqadUq83rsWLOeTFCQvXUB8E6EEQAN5OZK110nlZZKnTtL8+dLl19ud1UAvBljRgBIMrdlJk+Wzj/fBJEBA6R16wgiAFyvRWEkKytLvXv3VnBwsBISErRmzZqDHr97926NGTNGPXr0UFBQkPr06aPly5e3qGAArW/7dumcc8w6MpKZVXX1aun44+2tC4BvcPo2TU5OjtLS0pSdna2EhARlZmYqOTlZmzdvVvfu3Q84vrq6Wueee666d++uxYsXKyoqSlu2bFFYWFhr1A/gML37rnTttdKuXea2zLx5UkqK3VUB8CV+lmVZzpyQkJCggQMHavbs2ZIkh8Oh6OhojRs3ThMnTjzg+OzsbD366KPatGmT2rdwnujy8nKFhoaqrKxMISEhLfoMAA3V1Ej33Wd6QyzLPKr72mvSX/5id2UAvEVzv7+duk1TXV2tgoICJSUl1X+Av7+SkpKUn5/f6DlvvPGGEhMTNWbMGEVEROikk07SQw89pNra2iZ/TlVVlcrLyxtsAFrPtm3mtsyDD5og8s9/Svn5BBEA9nAqjJSWlqq2tlYREREN9kdERKioqKjRc3744QctXrxYtbW1Wr58uaZMmaLHH39cDzzwQJM/JyMjQ6GhoXVbdHS0M2UCOIi33za9IP/9r7kt8+qrUna2FBxsd2UAfJXLn6ZxOBzq3r27/vWvfykuLk4pKSmaPHmysrOzmzxn0qRJKisrq9u2bt3q6jIBr7dvn3T33dIFF5iVdgcMkAoKpCuvtLsyAL7OqQGs4eHhCggIUHFxcYP9xcXFioyMbPScHj16qH379goICKjb99e//lVFRUWqrq5WYGDgAecEBQUpiNmVgFZTWGgmMVu92rweM8ZMYkZvCAB34FTPSGBgoOLi4pSXl1e3z+FwKC8vT4mJiY2eM3jwYH333XdyOBx1+7755hv16NGj0SACoHW9+abpBVm9WgoJkRYtkmbPJogAcB9O36ZJS0vTvHnz9MILL+jrr7/WLbfcosrKSqWmpkqSRo4cqUmTJtUdf8stt+iXX37R7bffrm+++UbLli3TQw89pDFjxrTebwHgANXV0h13SBddJP3yixQXZyYxu+wyuysDgIacnmckJSVFu3btUnp6uoqKihQbG6vc3Ny6Qa2FhYXy96/PONHR0XrnnXc0YcIE9e/fX1FRUbr99tt19913t95vAaCBykpp+HBpfyfmbbdJjzzC2jIA3JPT84zYgXlGgOYrK5OGDpU+/ljq1El68UXp4ovtrgqAL2ru9zcL5QFe5JdfpORk6bPPpLAws+hdQoLdVQHAwRFGAC9RUiIlJUlffimFh5tp3gcMsLsqADg0wgjgBbZvl4YMkTZvliIjpfffl0480e6qAKB5CCOAh/vpJxNEfvhBio42g1aZ1h2AJ3H5DKwAXOfbb6UzzjBB5NhjpY8+IogA8DyEEcBDbdxogsjWrVLfviaI9O5td1UA4DzCCOCBPv9cOvNMqahI6t9fWrlSioqyuyoAaBnCCOBhcnNNEPn5Zyk+XvrwQ6l7d7urAoCWI4wAHiQry0xoVlFhAsn770tdu9pdFQAcHsII4AFqa6Xbb5fGjpUcDun66808IqGhdlcGAIePR3sBN1dRIV15pbR8uXmdkSHdfbfk52dvXQDQWggjgBsrLJQuvNDMqhocLL30EqvuAvA+hBHATa1ZI110kVRcbGZVfeMNaeBAu6sCgNbHmBHADS1ebAaoFhebR3c//ZQgAsB7EUYAN2JZ0owZ0uWXS3v3midnVq2SevWyuzIAcB3CCOBGHn5YmjTJtG+/XfrPf6TOne2tCQBcjTACuIm5c+uDyMMPS5mZUkCArSUBQJsgjABuICdHuuUW0548WbrrLnvrAYC2RBgBbPb229K115rxIrfcIt1/v90VAUDbIowANlq1Srr0UqmmRrrqKmn2bCYzA+B7CCOATdavNxOa7dkjXXCB9MILkj//IgH4IP7TB9jg22+l5GSprEw6/XRp0SKpfXu7qwIAexBGgDa2bZt07rlSSYkUGyu9+abUsaPdVQGAfQgjQBsqLZXOO0/askXq00d65x1W3gUAwgjQRioqpPPPl77+WjrqKOndd6Xu3e2uCgDsRxgB2kBVlTRihPTZZ1J4uPTee9LRR9tdFQC4B8II4GK1tWYekQ8+kDp1MvOK9O1rd1UA4D4II4ALWZY0dqxZhTcwUHr9dSk+3u6qAMC9EEYAF5o2TcrONhOZvfyyNGSI3RUBgPshjAAukpVlwsj+9uWX21sPALgrwgjgAjk50rhxpj1tWv0ieACAAxFGgFb23nvSddeZ8SJjxkhTpthdEQC4N8II0IrWrpUuvljat09KSZGeeoqF7wDgUAgjQCvZtMlMalZZaaZ7f/FFFr4DgObgP5VAK9i+3Sx89/PP0sCB0r//bR7lBQAcWovCSFZWlnr37q3g4GAlJCRozZo1TR77/PPPy8/Pr8EWHBzc4oIBd1NeLl1wgVRYKJ1wgrRsmdS5s91VAYDncDqM5OTkKC0tTVOnTtW6desUExOj5ORklZSUNHlOSEiIdu7cWbdt2bLlsIoG3EVNjRkbsmGDFBlpFr7r1s3uqgDAszgdRmbOnKnRo0crNTVV/fr1U3Z2tjp27Khnn322yXP8/PwUGRlZt0VERBxW0YA72D+7am6u1LGj9OabrDcDAC3hVBiprq5WQUGBkpKS6j/A319JSUnKz89v8rzffvtNRx99tKKjozV8+HBt3LjxoD+nqqpK5eXlDTbA3Tz2mDR3rnlaZsECpnkHgJZyKoyUlpaqtrb2gJ6NiIgIFRUVNXrOCSecoGeffVb/+c9/9PLLL8vhcOi0007Ttm3bmvw5GRkZCg0Nrduio6OdKRNwuUWLpLvuMu3MTGn4cFvLAQCP5vKnaRITEzVy5EjFxsbqzDPP1JIlS9StWzfNnTu3yXMmTZqksrKyum3r1q2uLhNottWrzaRmknTbbWYDALRcO2cODg8PV0BAgIqLixvsLy4uVmRkZLM+o3379howYIC+++67Jo8JCgpSUFCQM6UBbeL7700vSFWVdNFF0syZdlcEAJ7PqZ6RwMBAxcXFKS8vr26fw+FQXl6eEhMTm/UZtbW1+vLLL9WjRw/nKgVs9vPP5hHe0lIpLs6MEwkIsLsqAPB8TvWMSFJaWppGjRql+Ph4DRo0SJmZmaqsrFRqaqokaeTIkYqKilJGRoYkafr06Tr11FN1/PHHa/fu3Xr00Ue1ZcsW3Xjjja37mwAuVFVlpnn/5hupVy/z5MwRR9hdFQB4B6fDSEpKinbt2qX09HQVFRUpNjZWubm5dYNaCwsL5f+HObB//fVXjR49WkVFRerSpYvi4uK0evVq9evXr/V+C8CFLEv6xz+k//5XCgmRli+X6NgDgNbjZ1mWZXcRh1JeXq7Q0FCVlZUpJCTE7nLgY6ZPl6ZOldq1k95+W/rDk+0AgINo7vc3a9MAB7FqlTRtmmlnZxNEAMAVCCNAE8rLzSO8Doc0cqR0ww12VwQA3okwAjRh3Djpp5+k3r2lWbPsrgYAvBdhBGjEa69JL74o+ftLL79sBq4CAFyDMAL8ybZt0s03m/Y990iDB9tbDwB4O8II8AcOhzRqlPTrr2bhu/R0uysCAO9HGAH+4IknpA8+kDp2lF55RWrf3u6KAMD7EUaA//PFF+a2jGRCSZ8+9tYDAL6CMAJI2rtXuuYaqbpaGjZMGj3a7ooAwHcQRgBJEydKGzdK3btLzzwj+fnZXREA+A7CCHzeu+9KTz5p2s89ZwIJAKDtEEbg00pLpeuvN+1bb5UuuMDWcgDAJxFG4LMsS7rpJmnnTqlvX+nRR+2uCAB8E2EEPmvOHGnpUrMa7yuvmMd5AQBtjzACn7R+vZSWZtoPPyydcoqt5QCATyOMwOf89puUkiJVVUlDh0oTJthdEQD4NsIIfM6YMdI330hRUdLzz/MYLwDYjTACn/Lii/Wr8S5YIIWH210RAIAwAp+xebN5fFeS7rtPOuMMW8sBAPwfwgh8wp490hVXSJWV0jnn1K9BAwCwH2EEPuGOO6QNG6Ru3aSXX5YCAuyuCACwH2EEXu/f/zZzikhmvEiPHvbWAwBoiDACr/bjj9INN5j23XdLf/+7vfUAAA5EGIHX2rdPuuoqqaxMSkyU7r/f7ooAAI0hjMBrTZ4sffqpFBYmvfqq1L693RUBABpDGIFXWrWqfuG7+fOlo4+2tx4AQNMII/A6+/ZJN99s2v/4h3TJJfbWAwA4OMIIvM7MmdLGjWZ21UcesbsaAMChEEbgVX76SZo2zbQff1w68khbywEANANhBF7DsqSxY81sq2edJV13nd0VAQCagzACr7F0qbRsmXlqZs4cVuMFAE9BGIFXqKiQbrvNtO++W+rb1956AADNRxiBV0hPl7Zvl447jkXwAMDTEEbg8datk556yrSzsqQOHeytBwDgHMIIPFptrZlTxOGQUlKk5GS7KwIAOIswAo+WnS2tXSuFhEhPPGF3NQCAlmhRGMnKylLv3r0VHByshIQErVmzplnnLVy4UH5+fhoxYkRLfizQwM6d9eNDHnpI6tHD3noAAC3jdBjJyclRWlqapk6dqnXr1ikmJkbJyckqKSk56Hk//fST7rzzTp1++uktLhb4owkTpPJyKT6+fvp3AIDncTqMzJw5U6NHj1Zqaqr69eun7OxsdezYUc8++2yT59TW1uqaa67RtGnTdOyxxx5WwYAkvfOOlJMj+ftLc+dKAQF2VwQAaCmnwkh1dbUKCgqUlJRU/wH+/kpKSlJ+fn6T502fPl3du3fXDTfc0KyfU1VVpfLy8gYbsN+ePdKYMaY9bpx0yin21gMAODxOhZHS0lLV1tYqIiKiwf6IiAgVFRU1es6qVas0f/58zZs3r9k/JyMjQ6GhoXVbdHS0M2XCyz3yiPT991JUlHT//XZXAwA4XC59mqaiokLXXXed5s2bp/Dw8GafN2nSJJWVldVtW7dudWGV8CRbtkgzZpj2449LnTvbWw8A4PC1c+bg8PBwBQQEqLi4uMH+4uJiRUZGHnD8999/r59++knDhg2r2+dwOMwPbtdOmzdv1nHHHXfAeUFBQQoKCnKmNPiIO++U9u6VzjxTuuIKu6sBALQGp3pGAgMDFRcXp7y8vLp9DodDeXl5SkxMPOD4vn376ssvv9T69evrtosuukhnn3221q9fz+0XOOWDD6TFi82g1aeeYiE8APAWTvWMSFJaWppGjRql+Ph4DRo0SJmZmaqsrFRqaqokaeTIkYqKilJGRoaCg4N10kknNTg/LCxMkg7YDxxMTU39Qni33CL1729vPQCA1uN0GElJSdGuXbuUnp6uoqIixcbGKjc3t25Qa2Fhofz9mdgVrWvOHGnjRqlrV2n6dLurAQC0Jj/Lsiy7iziU8vJyhYaGqqysTCEhIXaXgza2a5fUp4+0e7cJJUxwBgCeobnf33RhwO1NnmyCSGysNHq03dUAAFobYQRuraBAeuYZ0541i5lWAcAbEUbgtizLzLBqWdLVV0t/+5vdFQEAXIEwArf1yitSfr50xBFm1lUAgHcijMAtVVRId91l2pMnm6nfAQDeiTACt/Tgg9LOndJxx0kTJthdDQDAlQgjcDvffivNnGnaTzwhBQfbWw8AwLUII3A748dL+/ZJf/+7dOGFdlcDAHA1wgjcyrJl0vLlUvv2UmYm688AgC8gjMBt1NZKd99t2rffLp1wgr31AADaBmEEbuPll836M2Fh0j332F0NAKCtEEbgFvbuldLTTXvSJKlLF3vrAQC0HcII3EJ2tlRYKPXsKY0da3c1AIC2RBiB7crLzbwiknTffVLHjraWAwBoY4QR2O7xx6XSUqlPHyk11e5qAABtjTACWxUXmzAimd6Rdu3srQcA0PYII7DVAw9IlZXSwIHSpZfaXQ0AwA6EEdjmhx+kuXNNe8YMJjgDAF9FGIFt0tPNtO/nniudc47d1QAA7EIYgS2++EJasMC0MzLsrQUAYC/CCGxxzz2SZUlXXCHFxdldDQDAToQRtLmPPjKL4bVrZwawAgB8G2EEbcqypIkTTfvGG6W//MXeegAA9iOMoE298YaUny916CBNmWJ3NQAAd0AYQZupra1fjXf8eLMODQAAhBG0mZdekr76yqzIe9dddlcDAHAXhBG0id9/r78tM2mSFBZmazkAADdCGEGbeOIJads2qVcvadw4u6sBALgTwghcrqiofmKzGTOk4GB76wEAuBfCCFwuPd0shjdokHTllXZXAwBwN4QRuNSXX0rz55v2zJkshgcAOBBhBC51552SwyFddpk0eLDd1QAA3BFhBC6Tmyu9+67Uvr0ZKwIAQGMII3CJmhrpjjtM+7bbpOOOs7ceAID7IozAJebPNxOcde0qTZ5sdzUAAHfWojCSlZWl3r17Kzg4WAkJCVqzZk2Txy5ZskTx8fEKCwvTEUccodjYWL300kstLhjur7zcPEEjSffdZ2ZcBQCgKU6HkZycHKWlpWnq1Klat26dYmJilJycrJKSkkaP79q1qyZPnqz8/Hxt2LBBqampSk1N1TvvvHPYxcM9zZghlZRIffpIN99sdzUAAHfnZ1mW5cwJCQkJGjhwoGbPni1Jcjgcio6O1rhx4zRx/9rwh3DKKado6NChuv/++5t1fHl5uUJDQ1VWVqaQkBBnykUb27JFOuEEqapKev11afhwuysCANilud/fTvWMVFdXq6CgQElJSfUf4O+vpKQk5efnH/J8y7KUl5enzZs364wzzmjyuKqqKpWXlzfY4BnuuccEkbPOki66yO5qAACewKkwUlpaqtraWkVERDTYHxERoaKioibPKysrU6dOnRQYGKihQ4dq1qxZOvfcc5s8PiMjQ6GhoXVbdHS0M2XCJmvWSAsWmInNHn+cCc4AAM3TJk/TdO7cWevXr9fatWv14IMPKi0tTStWrGjy+EmTJqmsrKxu27p1a1uUicNgWfWP8o4cKZ1yir31AAA8RztnDg4PD1dAQICKi4sb7C8uLlZkZGST5/n7++v444+XJMXGxurrr79WRkaGzjrrrEaPDwoKUlBQkDOlwWZLlkirVkkdOkgPPmh3NQAAT+JUz0hgYKDi4uKUl5dXt8/hcCgvL0+JiYnN/hyHw6GqqipnfjTc2N690l13mfadd0pRUfbWAwDwLE71jEhSWlqaRo0apfj4eA0aNEiZmZmqrKxUamqqJGnkyJGKiopSxv+tGZ+RkaH4+Hgdd9xxqqqq0vLly/XSSy9pzpw5rfubwDaPPy798IPUs2d9KAEAoLmcDiMpKSnatWuX0tPTVVRUpNjYWOXm5tYNai0sLJS/f32HS2VlpW699VZt27ZNHTp0UN++ffXyyy8rJSWl9X4L2KawsP62zGOPSZ062VsPAMDzOD3PiB2YZ8R9XX65tHixdMYZ0ooVPEEDAKjnknlGgD96/30TRAICpFmzCCIAgJYhjKBFqqulceNMe8wYqX9/e+sBAHguwgha5KmnpE2bpG7dpGnT7K4GAODJCCNw2o4d9QHk4YelsDBbywEAeDjCCJx2113Sb79JCQnSqFF2VwMA8HSEETjlo4+kV14xg1Vnz5b8+RsEADhMfJWg2Wpq6get3nSTFB9vbz0AAO9AGEGzZWdLGzZIXbuy/gwAoPUQRtAsJSXSlCmm/eCD0pFH2lsPAMB7EEbQLPfcI+3eLQ0YII0ebXc1AABvQhjBIX36qTR/vmnPnm1mXAUAoLUQRnBQtbXS2LGmPWqUdNpp9tYDAPA+hBEc1BNPSJ99JoWESDNm2F0NAMAbEUbQpI0bpcmTTfvxx6XISHvrAQB4J8IIGrVvnzRypFkQb+hQ6YYb7K4IAOCtCCNo1IMPSuvWmTlF5s0zM64CAOAKhBEc4LPPpAceMO2nn5Z69LC3HgCAdyOMoIE9e8ztmdpa6YorpJQUuysCAHg7wggamDJF+vprM1j16aftrgYA4AsII6jz0UfSzJmmPW8eU74DANoGYQSSpIoK6frrJcuS/vEP6cIL7a4IAOArCCOQJN15p/Tjj9LRR5uJzgAAaCuEESg3V/rXv0z7uefMbKsAALQVwoiP+/XX+gnNbrtNOvtse+sBAPgewoiPGztW2rFD6tNHysiwuxoAgC8ijPiwxYulBQskf3/pxReljh3trggA4IsIIz5qxw7pn/807UmTpIQEe+sBAPguwogP2v/47i+/SKecIqWn210RAMCXEUZ80NNPS++8IwUHSy+/LAUG2l0RAMCXEUZ8zKZNZk4RSXrkEemvf7W3HgAACCM+ZN8+6dprpb17pXPPlcaMsbsiAAAIIz5l+nSpoEDq0sVMbubP//oAADfA15GPyM+XHnrItLOzpagoe+sBAGA/wogP+O036brrJIfD3Ka54gq7KwIAoB5hxAekpUnffy9FR0uzZtldDQAADRFGvNybb0rz5kl+ftILL0hhYXZXBABAQy0KI1lZWerdu7eCg4OVkJCgNWvWNHnsvHnzdPrpp6tLly7q0qWLkpKSDno8Wk9JiXTjjaadlsYieAAA9+R0GMnJyVFaWpqmTp2qdevWKSYmRsnJySopKWn0+BUrVuiqq67Shx9+qPz8fEVHR+u8887T9u3bD7t4NM2ypNGjTSA5+WTpgQfsrggAgMb5WZZlOXNCQkKCBg4cqNmzZ0uSHA6HoqOjNW7cOE2cOPGQ59fW1qpLly6aPXu2Ro4c2ayfWV5ertDQUJWVlSkkJMSZcn1WZqY0YYKZXXXtWql/f7srAgD4muZ+fzvVM1JdXa2CggIlJSXVf4C/v5KSkpSfn9+sz/j999+1b98+de3atcljqqqqVF5e3mBD882da4KIZB7nJYgAANyZU2GktLRUtbW1ioiIaLA/IiJCRUVFzfqMu+++Wz179mwQaP4sIyNDoaGhdVt0dLQzZfq0+fOlm2827TvvNGNFAABwZ236NM2MGTO0cOFCLV26VMHBwU0eN2nSJJWVldVtW7dubcMqPdcLL5hxIpI0frxZe8bPz9aSAAA4pHbOHBweHq6AgAAVFxc32F9cXKzIyMiDnvvYY49pxowZev/999X/EPcNgoKCFBQU5ExpPu/ll6XUVDNwdexYaeZMgggAwDM41TMSGBiouLg45eXl1e1zOBzKy8tTYmJik+c98sgjuv/++5Wbm6v4+PiWV4tGLVwojRplgsjNN0tPPUUQAQB4Dqd6RiQpLS1No0aNUnx8vAYNGqTMzExVVlYqNTVVkjRy5EhFRUUpIyNDkvTwww8rPT1dCxYsUO/evevGlnTq1EmdOnVqxV/FNy1ebKZ4dzjMnCJZWQQRAIBncTqMpKSkaNeuXUpPT1dRUZFiY2OVm5tbN6i1sLBQ/n9YDnbOnDmqrq7WZZdd1uBzpk6dqvvuu+/wqvdxr78uXXWVVFsrXX+9eYqGlXgBAJ7G6XlG7MA8Iwd6803p0kulfftMz8jzz0sBAXZXBQBAPZfMMwL38Pbb0mWXmSBy5ZXSc88RRAAAnosw4mHee0+6+GKputoEkpdekto5fbMNAAD3QRjxICtWSMOHS1VV5s8FCwgiAADPRxjxEKtWSRdeKO3ZIw0dKuXkSO3b210VAACHjzDiAT79VLrgAqmyUjrvPPM4L3PCAQC8BWHEzRUUSMnJUkWFdPbZ0tKl0kFm0gcAwOMQRtzYF19I554rlZVJf/ubeZy3Y0e7qwIAoHURRtzUxo1SUpL066/SqadKy5dLRxxhd1UAALQ+wogb2rRJGjJEKi2V4uLMvCKdO9tdFQAArkEYcTPffSedc45UXCzFxEjvviuFhdldFQAArkMYcSObN5tBqjt3SiedJL3/vtS1q91VAQDgWoQRN/HFF9Lpp0vbtkl//asJIuHhdlcFAIDrEUbcwCefSGedJe3aJQ0YIK1cKf3fIsgAAHg9wojNPvjAPDWze7d02mnmdbdudlcFAEDbIYzYaNmy+plVhwxhsCoAwDcRRmzy2mvSiBFm0buLLpLeeot5RAAAvokwYoNnn5WuukqqqZGuvtqsNcMU7wAAX0UYaWNPPindcIPkcEg33SS9+CKr7wIAfBthpA09+KA0frxp33GHlJ0tBQTYWhIAALYjjLSRuXOle+817fvukx59VPLzs7UkAADcQju7C/AFK1ZIY8ea9n33SVOn2lkNAADuhZ4RF/vhB+nSS81g1SuvlNLT7a4IAAD3QhhxofJyadgw6ZdfpPh48xQNt2YAAGiIMOIitbXmsd2vvpJ69JBef13q0MHuqgAAcD+EEReZNMnMsBocbIJIVJTdFQEA4J4IIy7wwgvmaRnJ3JoZNMjeegAAcGeEkVa2erWZzEySJk82M60CAICmEUZaUWGhdPHFUnW1WXdm+nS7KwIAwP0RRlrJb7+ZBe9KSqT+/aWXXpL8uboAABwSX5etwOGQRo2SvvhC6tZNeuMNqVMnu6sCAMAzEEZaQWamtGSJWfBu6VLp6KPtrggAAM9BGDlMX38t3XOPaT/5pDR4sL31AADgaQgjh6GmxtyeqaqSkpOlm2+2uyIAADwPYeQwPPywtHatFBYmzZ/PVO8AALQEYaSF1q+Xpk0z7VmzmGEVAICWalEYycrKUu/evRUcHKyEhAStWbOmyWM3btyoSy+9VL1795afn58yMzNbWqvbqKoyt2f27TPziVxzjd0VAQDguZwOIzk5OUpLS9PUqVO1bt06xcTEKDk5WSUlJY0e//vvv+vYY4/VjBkzFBkZedgFu4Pp06UNG6TwcCk7m9szAAAcDqfDyMyZMzV69GilpqaqX79+ys7OVseOHfXss882evzAgQP16KOP6sorr1RQUNBhF2y3Tz+VZsww7TlzpIgIe+sBAMDTORVGqqurVVBQoKSkpPoP8PdXUlKS8vPzW62oqqoqlZeXN9jcwZ495vaMwyFdfbV02WV2VwQAgOdzKoyUlpaqtrZWEX/qDoiIiFBRUVGrFZWRkaHQ0NC6LTo6utU++3BMnixt3iz16GEGrQIAgMPnlk/TTJo0SWVlZXXb1q1b7S5JK1eamVYl6ZlnpK5dbS0HAACv0c6Zg8PDwxUQEKDi4uIG+4uLi1t1cGpQUJBbjS+pqJBSUyXLkm64QbrgArsrAgDAezjVMxIYGKi4uDjl5eXV7XM4HMrLy1NiYmKrF+cu/t//k378UerVS5o50+5qAADwLk71jEhSWlqaRo0apfj4eA0aNEiZmZmqrKxUamqqJGnkyJGKiopSRkaGJDPo9auvvqprb9++XevXr1enTp10/PHHt+Kv4hrvvivNnWvazz0nhYTYWw8AAN7G6TCSkpKiXbt2KT09XUVFRYqNjVVubm7doNbCwkL5+9d3uOzYsUMDBgyoe/3YY4/pscce05lnnqkVK1Yc/m/gQhUV0ujRpj12rHTOOfbWAwCAN/KzLMuyu4hDKS8vV2hoqMrKyhTShl0Tt91mnprp3Vv68kupU6c2+9EAAHi85n5/u+XTNO7g44+l2bNN+1//IogAAOAqhJFG7N0r3XijeXomNVU691y7KwIAwHsRRhrxwAPSpk1SZKT0+ON2VwMAgHcjjPzJF19IDz9s2llZUpcu9tYDAIC3I4z8QU2NmdSspka65BKzAQAA1yKM/METT0gFBVJYWP3gVQAA4FqEkf/z7bdSerppz5xpFsMDAACuRxiR5HCYyc327pWSkqTrr7e7IgAAfAdhRGYV3pUrpY4dzZwifn52VwQAgO/w+TCybZtZCE+SHnpIOuYYe+sBAMDX+HQYsSzplluk8nLp1FPN+jMAAKBt+XQYycmR3npLat/e3KoJCLC7IgAAfI/PhpE9e6Tx40373nulE0+0tRwAAHyWz4aRDh2kf/9buvRSaeJEu6sBAMB3tbO7ADsNHmw2AABgH5/tGQEAAO6BMAIAAGxFGAEAALYijAAAAFsRRgAAgK0IIwAAwFaEEQAAYCvCCAAAsBVhBAAA2IowAgAAbEUYAQAAtiKMAAAAWxFGAACArTxi1V7LsiRJ5eXlNlcCAACaa//39v7v8aZ4RBipqKiQJEVHR9tcCQAAcFZFRYVCQ0ObfN/POlRccQMOh0M7duxQ586d5efn12qfW15erujoaG3dulUhISGt9rloHNe7bXG92xbXu21xvdtWS6+3ZVmqqKhQz5495e/f9MgQj+gZ8ff311FHHeWyzw8JCeEvcxvierctrnfb4nq3La5322rJ9T5Yj8h+DGAFAAC2IowAAABb+XQYCQoK0tSpUxUUFGR3KT6B6922uN5ti+vdtrjebcvV19sjBrACAADv5dM9IwAAwH6EEQAAYCvCCAAAsBVhBAAA2Mqnw0hWVpZ69+6t4OBgJSQkaM2aNXaX5BU++ugjDRs2TD179pSfn59ef/31Bu9blqX09HT16NFDHTp0UFJSkr799lt7ivUCGRkZGjhwoDp37qzu3btrxIgR2rx5c4Nj9u7dqzFjxujII49Up06ddOmll6q4uNimij3bnDlz1L9//7rJnxITE/X222/Xvc+1dp0ZM2bIz89P48ePr9vH9W5d9913n/z8/Bpsffv2rXvfVdfbZ8NITk6O0tLSNHXqVK1bt04xMTFKTk5WSUmJ3aV5vMrKSsXExCgrK6vR9x955BE99dRTys7O1qeffqojjjhCycnJ2rt3bxtX6h1WrlypMWPG6JNPPtF7772nffv26bzzzlNlZWXdMRMmTNCbb76pRYsWaeXKldqxY4cuueQSG6v2XEcddZRmzJihgoICffbZZzrnnHM0fPhwbdy4URLX2lXWrl2ruXPnqn///g32c71b34knnqidO3fWbatWrap7z2XX2/JRgwYNssaMGVP3ura21urZs6eVkZFhY1XeR5K1dOnSutcOh8OKjIy0Hn300bp9u3fvtoKCgqxXX33Vhgq9T0lJiSXJWrlypWVZ5vq2b9/eWrRoUd0xX3/9tSXJys/Pt6tMr9KlSxfrmWee4Vq7SEVFhfWXv/zFeu+996wzzzzTuv322y3L4u+2K0ydOtWKiYlp9D1XXm+f7Bmprq5WQUGBkpKS6vb5+/srKSlJ+fn5Nlbm/X788UcVFRU1uPahoaFKSEjg2reSsrIySVLXrl0lSQUFBdq3b1+Da963b1/16tWLa36YamtrtXDhQlVWVioxMZFr7SJjxozR0KFDG1xXib/brvLtt9+qZ8+eOvbYY3XNNdeosLBQkmuvt0cslNfaSktLVVtbq4iIiAb7IyIitGnTJpuq8g1FRUWS1Oi13/8eWs7hcGj8+PEaPHiwTjrpJEnmmgcGBiosLKzBsVzzlvvyyy+VmJiovXv3qlOnTlq6dKn69eun9evXc61b2cKFC7Vu3TqtXbv2gPf4u936EhIS9Pzzz+uEE07Qzp07NW3aNJ1++un63//+59Lr7ZNhBPBWY8aM0f/+978G93jR+k444QStX79eZWVlWrx4sUaNGqWVK1faXZbX2bp1q26//Xa99957Cg4Otrscn3D++efXtfv376+EhAQdffTReu2119ShQweX/VyfvE0THh6ugICAA0YAFxcXKzIy0qaqfMP+68u1b31jx47VW2+9pQ8//FBHHXVU3f7IyEhVV1dr9+7dDY7nmrdcYGCgjj/+eMXFxSkjI0MxMTF68sknudatrKCgQCUlJTrllFPUrl07tWvXTitXrtRTTz2ldu3aKSIiguvtYmFhYerTp4++++47l/799skwEhgYqLi4OOXl5dXtczgcysvLU2Jioo2Veb9jjjlGkZGRDa59eXm5Pv30U659C1mWpbFjx2rp0qX64IMPdMwxxzR4Py4uTu3bt29wzTdv3qzCwkKueStxOByqqqriWreyIUOG6Msvv9T69evrtvj4eF1zzTV1ba63a/3222/6/vvv1aNHD9f+/T6s4a8ebOHChVZQUJD1/PPPW1999ZV10003WWFhYVZRUZHdpXm8iooK6/PPP7c+//xzS5I1c+ZM6/PPP7e2bNliWZZlzZgxwwoLC7P+85//WBs2bLCGDx9uHXPMMdaePXtsrtwz3XLLLVZoaKi1YsUKa+fOnXXb77//XnfMzTffbPXq1cv64IMPrM8++8xKTEy0EhMTbazac02cONFauXKl9eOPP1obNmywJk6caPn5+VnvvvuuZVlca1f749M0lsX1bm133HGHtWLFCuvHH3+0Pv74YyspKckKDw+3SkpKLMty3fX22TBiWZY1a9Ysq1evXlZgYKA1aNAg65NPPrG7JK/w4YcfWpIO2EaNGmVZlnm8d8qUKVZERIQVFBRkDRkyxNq8ebO9RXuwxq61JOu5556rO2bPnj3WrbfeanXp0sXq2LGjdfHFF1s7d+60r2gP9o9//MM6+uijrcDAQKtbt27WkCFD6oKIZXGtXe3PYYTr3bpSUlKsHj16WIGBgVZUVJSVkpJifffdd3Xvu+p6+1mWZR1e3woAAEDL+eSYEQAA4D4IIwAAwFaEEQAAYCvCCAAAsBVhBAAA2IowAgAAbEUYAQAAtiKMAAAAWxFGAACArQgjAADAVoQRAABgK8IIAACw1f8HOmn1gy7+b5QAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot the accuracy and loss\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "acc = history.history['accuracy']\n",
    "loss = history.history['loss']\n",
    "epochs = range(len(acc))\n",
    "\n",
    "plt.plot(epochs, acc, 'b', label='Training accuracy')\n",
    "plt.figure()\n",
    "plt.plot(epochs, loss, 'r', label='Training Loss')\n",
    "plt.title('Training accuracy and loss')\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: model\\assets\n"
     ]
    }
   ],
   "source": [
    "# Saving the model and tokenizer\n",
    "history.model.save('model', save_format='tf')\n",
    "tokenizer_json = tokenizer.to_json()\n",
    "with open('tokenizer.json', 'w', encoding='utf-8') as f:\n",
    "    f.write(tokenizer_json)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": ".venv",
   "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": "3.10.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}