LSTM/next_word_predictor.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# LSTM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(\"./data/next_word_prediction.txt\", encoding=\"utf-8\") as f:\n",
    "    text = f.read()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Nepal is a landlocked country with highly diverse and rich geography, culture, and religions. It is '"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "text[:100]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "from tensorflow.keras.preprocessing.text import Tokenizer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "tokenizer = Tokenizer(num_words=1000, oov_token=\"<UNK>\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "tokenizer.fit_on_texts([text])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "word2index = tokenizer.word_index\n",
    "\n",
    "index2word = {v: k for k, v  in word2index.items()}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['Nepal is a landlocked country with highly diverse and rich geography, culture, and religions',\n",
       " ' It is situated in South Asia on the lap of the huge Himalayas surrounded by India in the east, south, and west and China in the north',\n",
       " ' It occupies an area of 147,516 sq',\n",
       " ' km',\n",
       " ' with an average of 193km wide from North to South and an average length (East to West) is 885km',\n",
       " ' It provides a variety of geographical structures owning a large number of natural gifts',\n",
       " '\\n\\nGeographically, Nepal is divided into three regions',\n",
       " ' The mountain regions, Hilly and Terai regions occupy about 15%, 68%, and 17% of the total area of Nepal respectively',\n",
       " ' Nepal owns many snow-capped lofty mountains',\n",
       " ' The highest peak in the world, Mt']"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "text_split = text.split(\".\")\n",
    "\n",
    "text_split[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "text_sequences = list()\n",
    "\n",
    "for sentence in text.split(\".\"):\n",
    "    tokenized = tokenizer.texts_to_sequences([sentence])[0]\n",
    "    text_sequences.append(tokenized)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[30, 120, 121, 3, 48, 122, 123]"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "text_sequences[10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "decode = lambda x: \" \".join([index2word[i] for i in x])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[5, 7, 6, 1, 22]\n",
      "nepal is a <UNK> country\n"
     ]
    }
   ],
   "source": [
    "text_sample = \"Nepal is a beautiful country\"\n",
    "\n",
    "encoded = tokenizer.texts_to_sequences([text_sample])[0]\n",
    "print(encoded)\n",
    "decoded = decode(encoded)\n",
    "print(decoded)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "sample_data = [[1,2,3],[1,2],[1,2,3,4,5,6,7,8,9,0]]\n",
    "sample_train_data = list()\n",
    "\n",
    "for data in sample_data:\n",
    "    for i in range(1, len(data)):\n",
    "        input = data[:i+1]\n",
    "        sample_train_data.append(input)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[[1, 2],\n",
       " [1, 2, 3],\n",
       " [1, 2],\n",
       " [1, 2],\n",
       " [1, 2, 3],\n",
       " [1, 2, 3, 4],\n",
       " [1, 2, 3, 4, 5],\n",
       " [1, 2, 3, 4, 5, 6],\n",
       " [1, 2, 3, 4, 5, 6, 7],\n",
       " [1, 2, 3, 4, 5, 6, 7, 8],\n",
       " [1, 2, 3, 4, 5, 6, 7, 8, 9],\n",
       " [1, 2, 3, 4, 5, 6, 7, 8, 9, 0]]"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sample_train_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Data: [1] -> 2\n",
      "Data: [1, 2] -> 3\n",
      "Data: [1] -> 2\n",
      "Data: [1] -> 2\n",
      "Data: [1, 2] -> 3\n",
      "Data: [1, 2, 3] -> 4\n",
      "Data: [1, 2, 3, 4] -> 5\n",
      "Data: [1, 2, 3, 4, 5] -> 6\n",
      "Data: [1, 2, 3, 4, 5, 6] -> 7\n",
      "Data: [1, 2, 3, 4, 5, 6, 7] -> 8\n",
      "Data: [1, 2, 3, 4, 5, 6, 7, 8] -> 9\n",
      "Data: [1, 2, 3, 4, 5, 6, 7, 8, 9] -> 0\n"
     ]
    }
   ],
   "source": [
    "for seq in sample_train_data:\n",
    "    sample_data = seq[:-1]\n",
    "    sample_target = seq[-1]\n",
    "    \n",
    "    print(f\"Data: {sample_data} -> {sample_target}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_sequences = list()\n",
    "\n",
    "for sequence in text_sequences:\n",
    "    for i in range(1, len(sequence)):\n",
    "        train_sequences.append(sequence[:i+1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_sequence_features = [seq[:-1] for seq in train_sequences]\n",
    "train_sequence_labels = [seq[-1] for seq in train_sequences]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Data: [5] -> 7\n",
      "Data: [5, 7] -> 6\n",
      "Data: [5, 7, 6] -> 74\n",
      "Data: [5, 7, 6, 74] -> 22\n",
      "Data: [5, 7, 6, 74, 22] -> 23\n",
      "Data: [5, 7, 6, 74, 22, 23] -> 75\n",
      "Data: [5, 7, 6, 74, 22, 23, 75] -> 76\n",
      "Data: [5, 7, 6, 74, 22, 23, 75, 76] -> 4\n",
      "Data: [5, 7, 6, 74, 22, 23, 75, 76, 4] -> 35\n",
      "Data: [5, 7, 6, 74, 22, 23, 75, 76, 4, 35] -> 77\n",
      "Data: [5, 7, 6, 74, 22, 23, 75, 76, 4, 35, 77] -> 78\n",
      "Data: [5, 7, 6, 74, 22, 23, 75, 76, 4, 35, 77, 78] -> 4\n",
      "Data: [5, 7, 6, 74, 22, 23, 75, 76, 4, 35, 77, 78, 4] -> 79\n",
      "Data: [9] -> 7\n",
      "Data: [9, 7] -> 80\n"
     ]
    }
   ],
   "source": [
    "for f, t in zip(train_sequence_features[:15], train_sequence_labels[:15]):\n",
    "    print(f\"Data: {f} -> {t}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "30\n"
     ]
    }
   ],
   "source": [
    "max_sequence_length = max([len(seq) for seq in train_sequence_features])\n",
    "\n",
    "print(max_sequence_length)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[5]"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_sequence_features[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "def pad_sequence(sequence, max_sequence_length):\n",
    "    \"\"\"Takes a sequence and return a 0 pre padded sequence of max length sequence\"\"\"\n",
    "    padded_sequence = list()\n",
    "    zeros_required = max_sequence_length - len(sequence)\n",
    "    padded_sequence.append([0]*zeros_required + sequence)\n",
    "    return padded_sequence[0]\n",
    "    \n",
    "    \n",
    "padded_sequence = pad_sequence([1,2,3], 5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[0, 0, 1, 2, 3]"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "padded_sequence"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "padded_train_sequences = [pad_sequence(seq, max_sequence_length) for seq in train_sequence_features]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 5, 7, 6, 74, 22, 23, 75, 76, 4, 35, 77] 78\n"
     ]
    }
   ],
   "source": [
    "print(padded_train_sequences[10], train_sequence_labels[10])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "x_train = tf.constant(padded_train_sequences, dtype=tf.float32)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.utils import to_categorical\n",
    "\n",
    "y_train = to_categorical(train_sequence_labels, num_classes=len(word2index) + 1)\n",
    "\n",
    "y_train = tf.constant(y_train, dtype=tf.float32)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(TensorShape([475, 30]), TensorShape([475, 284]))"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_train.shape, y_train.shape # train feature is tokenized sequence of max sequence 30 and label shape output is one hot encoded label of vocab size 283 of our dataset text"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "# importing\n",
    "\n",
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Embedding, LSTM, Dense"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\acer\\AppData\\Roaming\\Python\\Python312\\site-packages\\keras\\src\\layers\\core\\embedding.py:90: UserWarning: Argument `input_length` is deprecated. Just remove it.\n",
      "  warnings.warn(\n"
     ]
    }
   ],
   "source": [
    "model = Sequential()\n",
    "\n",
    "model.add(Embedding(input_dim=len(word2index) + 1, output_dim=32, input_length=max_sequence_length))\n",
    "model.add(LSTM(150))\n",
    "model.add(Dense(len(word2index) + 1, activation=\"softmax\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\acer\\AppData\\Roaming\\Python\\Python312\\site-packages\\keras\\src\\layers\\core\\embedding.py:90: UserWarning: Argument `input_length` is deprecated. Just remove it.\n",
      "  warnings.warn(\n"
     ]
    }
   ],
   "source": [
    "from tensorflow.keras.models import Model\n",
    "from tensorflow.keras.layers import Input\n",
    "\n",
    "vocab_size = len(word2index) + 1\n",
    "\n",
    "input = Input(shape=(max_sequence_length,))\n",
    "x = Embedding(input_dim=vocab_size, output_dim=32, input_length=max_sequence_length)(input)\n",
    "x = LSTM(150)(x)\n",
    "output = Dense(vocab_size, activation=\"softmax\")(x)\n",
    "\n",
    "model = Model(inputs=input, outputs=output)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\">Model: \"functional_1\"</span>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1mModel: \"functional_1\"\u001b[0m\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓\n",
       "┃<span style=\"font-weight: bold\"> Layer (type)                    </span>┃<span style=\"font-weight: bold\"> Output Shape           </span>┃<span style=\"font-weight: bold\">       Param # </span>┃\n",
       "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩\n",
       "│ input_layer (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">InputLayer</span>)        │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">30</span>)             │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ embedding_1 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Embedding</span>)         │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">30</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>)         │         <span style=\"color: #00af00; text-decoration-color: #00af00\">9,088</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ lstm_1 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">LSTM</span>)                   │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">150</span>)            │       <span style=\"color: #00af00; text-decoration-color: #00af00\">109,800</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_1 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                 │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">284</span>)            │        <span style=\"color: #00af00; text-decoration-color: #00af00\">42,884</span> │\n",
       "└─────────────────────────────────┴────────────────────────┴───────────────┘\n",
       "</pre>\n"
      ],
      "text/plain": [
       "┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓\n",
       "┃\u001b[1m \u001b[0m\u001b[1mLayer (type)                   \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mOutput Shape          \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1m      Param #\u001b[0m\u001b[1m \u001b[0m┃\n",
       "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩\n",
       "│ input_layer (\u001b[38;5;33mInputLayer\u001b[0m)        │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m30\u001b[0m)             │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ embedding_1 (\u001b[38;5;33mEmbedding\u001b[0m)         │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m30\u001b[0m, \u001b[38;5;34m32\u001b[0m)         │         \u001b[38;5;34m9,088\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ lstm_1 (\u001b[38;5;33mLSTM\u001b[0m)                   │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m150\u001b[0m)            │       \u001b[38;5;34m109,800\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_1 (\u001b[38;5;33mDense\u001b[0m)                 │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m284\u001b[0m)            │        \u001b[38;5;34m42,884\u001b[0m │\n",
       "└─────────────────────────────────┴────────────────────────┴───────────────┘\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Total params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">161,772</span> (631.92 KB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Total params: \u001b[0m\u001b[38;5;34m161,772\u001b[0m (631.92 KB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">161,772</span> (631.92 KB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Trainable params: \u001b[0m\u001b[38;5;34m161,772\u001b[0m (631.92 KB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "text/html": [
       "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Non-trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> (0.00 B)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Non-trainable params: \u001b[0m\u001b[38;5;34m0\u001b[0m (0.00 B)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.compile(loss=\"categorical_crossentropy\", optimizer=\"adam\", metrics=[\"accuracy\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 14ms/step - accuracy: 0.0076 - loss: 5.6471\n",
      "Epoch 2/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.0406 - loss: 5.5046\n",
      "Epoch 3/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.0510 - loss: 5.2578\n",
      "Epoch 4/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.0605 - loss: 5.2167\n",
      "Epoch 5/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.0260 - loss: 5.2507\n",
      "Epoch 6/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.0571 - loss: 5.1840\n",
      "Epoch 7/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 17ms/step - accuracy: 0.0344 - loss: 5.1607\n",
      "Epoch 8/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.0578 - loss: 5.1278\n",
      "Epoch 9/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 14ms/step - accuracy: 0.0552 - loss: 5.1135\n",
      "Epoch 10/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.0637 - loss: 5.0853\n",
      "Epoch 11/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.0798 - loss: 4.9980\n",
      "Epoch 12/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.0616 - loss: 4.9800 \n",
      "Epoch 13/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 15ms/step - accuracy: 0.0675 - loss: 4.8985\n",
      "Epoch 14/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.0824 - loss: 4.7128\n",
      "Epoch 15/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.0731 - loss: 4.7600\n",
      "Epoch 16/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.0839 - loss: 4.6284\n",
      "Epoch 17/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 11ms/step - accuracy: 0.0871 - loss: 4.6080\n",
      "Epoch 18/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.0962 - loss: 4.4390\n",
      "Epoch 19/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 14ms/step - accuracy: 0.1205 - loss: 4.3555\n",
      "Epoch 20/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 11ms/step - accuracy: 0.1256 - loss: 4.2027\n",
      "Epoch 21/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.0971 - loss: 4.2282\n",
      "Epoch 22/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.1286 - loss: 4.0095\n",
      "Epoch 23/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 14ms/step - accuracy: 0.1488 - loss: 3.9126\n",
      "Epoch 24/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.1591 - loss: 3.8320\n",
      "Epoch 25/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.1723 - loss: 3.6332\n",
      "Epoch 26/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 18ms/step - accuracy: 0.1880 - loss: 3.5851\n",
      "Epoch 27/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 15ms/step - accuracy: 0.1905 - loss: 3.4070\n",
      "Epoch 28/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.2181 - loss: 3.2609\n",
      "Epoch 29/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.2642 - loss: 3.1508\n",
      "Epoch 30/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.3053 - loss: 2.9913\n",
      "Epoch 31/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 14ms/step - accuracy: 0.3224 - loss: 2.9070\n",
      "Epoch 32/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 15ms/step - accuracy: 0.3844 - loss: 2.8210\n",
      "Epoch 33/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.3866 - loss: 2.6462\n",
      "Epoch 34/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.4220 - loss: 2.5861\n",
      "Epoch 35/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.4917 - loss: 2.4189\n",
      "Epoch 36/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.5223 - loss: 2.3735\n",
      "Epoch 37/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.5498 - loss: 2.1848\n",
      "Epoch 38/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 14ms/step - accuracy: 0.5450 - loss: 2.1488\n",
      "Epoch 39/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.6021 - loss: 2.0452\n",
      "Epoch 40/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 14ms/step - accuracy: 0.6720 - loss: 1.9102\n",
      "Epoch 41/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 14ms/step - accuracy: 0.6824 - loss: 1.9007\n",
      "Epoch 42/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.7074 - loss: 1.8241\n",
      "Epoch 43/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 15ms/step - accuracy: 0.7364 - loss: 1.6704\n",
      "Epoch 44/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 14ms/step - accuracy: 0.7734 - loss: 1.6255\n",
      "Epoch 45/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 19ms/step - accuracy: 0.7797 - loss: 1.5635\n",
      "Epoch 46/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.7924 - loss: 1.5103\n",
      "Epoch 47/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.8410 - loss: 1.3844\n",
      "Epoch 48/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.8388 - loss: 1.3677\n",
      "Epoch 49/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 14ms/step - accuracy: 0.8364 - loss: 1.3114\n",
      "Epoch 50/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.8858 - loss: 1.2470\n",
      "Epoch 51/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.8869 - loss: 1.1528\n",
      "Epoch 52/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.8742 - loss: 1.1592\n",
      "Epoch 53/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.8890 - loss: 1.1025\n",
      "Epoch 54/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.9114 - loss: 1.0031\n",
      "Epoch 55/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.8815 - loss: 1.0381\n",
      "Epoch 56/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 16ms/step - accuracy: 0.8982 - loss: 1.0064\n",
      "Epoch 57/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 14ms/step - accuracy: 0.9247 - loss: 0.9106\n",
      "Epoch 58/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 14ms/step - accuracy: 0.9166 - loss: 0.8761\n",
      "Epoch 59/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 14ms/step - accuracy: 0.9238 - loss: 0.8339\n",
      "Epoch 60/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.9445 - loss: 0.7745\n",
      "Epoch 61/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.9200 - loss: 0.7997\n",
      "Epoch 62/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 14ms/step - accuracy: 0.9310 - loss: 0.7485\n",
      "Epoch 63/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 14ms/step - accuracy: 0.9153 - loss: 0.7800\n",
      "Epoch 64/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 15ms/step - accuracy: 0.9332 - loss: 0.7156\n",
      "Epoch 65/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.9298 - loss: 0.6876\n",
      "Epoch 66/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.9404 - loss: 0.6542\n",
      "Epoch 67/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.9515 - loss: 0.6310\n",
      "Epoch 68/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.9334 - loss: 0.6059\n",
      "Epoch 69/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.9518 - loss: 0.5708\n",
      "Epoch 70/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 14ms/step - accuracy: 0.9335 - loss: 0.5908\n",
      "Epoch 71/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.9602 - loss: 0.5019\n",
      "Epoch 72/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 17ms/step - accuracy: 0.9238 - loss: 0.5588\n",
      "Epoch 73/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 15ms/step - accuracy: 0.9448 - loss: 0.5162\n",
      "Epoch 74/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.9566 - loss: 0.4882\n",
      "Epoch 75/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.9515 - loss: 0.4497\n",
      "Epoch 76/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.9539 - loss: 0.4758\n",
      "Epoch 77/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.9607 - loss: 0.4177\n",
      "Epoch 78/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.9611 - loss: 0.4248\n",
      "Epoch 79/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.9664 - loss: 0.4134\n",
      "Epoch 80/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.9578 - loss: 0.4102\n",
      "Epoch 81/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.9508 - loss: 0.3992\n",
      "Epoch 82/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.9520 - loss: 0.4016\n",
      "Epoch 83/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.9566 - loss: 0.3598\n",
      "Epoch 84/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 17ms/step - accuracy: 0.9653 - loss: 0.3458\n",
      "Epoch 85/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.9644 - loss: 0.3491\n",
      "Epoch 86/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.9420 - loss: 0.3560\n",
      "Epoch 87/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.9593 - loss: 0.3220\n",
      "Epoch 88/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 14ms/step - accuracy: 0.9578 - loss: 0.3334\n",
      "Epoch 89/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.9551 - loss: 0.3235\n",
      "Epoch 90/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.9653 - loss: 0.2782\n",
      "Epoch 91/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.9578 - loss: 0.2998\n",
      "Epoch 92/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.9478 - loss: 0.2992\n",
      "Epoch 93/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.9682 - loss: 0.2670\n",
      "Epoch 94/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.9597 - loss: 0.2830\n",
      "Epoch 95/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.9492 - loss: 0.2752\n",
      "Epoch 96/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.9547 - loss: 0.2713\n",
      "Epoch 97/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 14ms/step - accuracy: 0.9524 - loss: 0.2564\n",
      "Epoch 98/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 13ms/step - accuracy: 0.9537 - loss: 0.2579\n",
      "Epoch 99/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.9642 - loss: 0.2307\n",
      "Epoch 100/100\n",
      "\u001b[1m15/15\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 12ms/step - accuracy: 0.9699 - loss: 0.2289\n"
     ]
    }
   ],
   "source": [
    "model_history = model.fit(x_train, y_train, epochs=100, verbose=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x1df72f07e00>]"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "plt.plot(model_history.history[\"accuracy\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Loss vs Epoch')"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAioAAAHHCAYAAACRAnNyAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjguMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8g+/7EAAAACXBIWXMAAA9hAAAPYQGoP6dpAABIqUlEQVR4nO3dd3gU1cIG8Hd2N9lN201vpBAgJBBI6AEpioQm0kQpoiJXRRAU7Be5KFa4Fi7XRlGBawNBpVykSJEmJZSEGkILEEgnbDZ1U/Z8f0T2Mzcg6bPl/T3PPJiZyfLmPN8l7zcz54wkhBAgIiIiskAKuQMQERER3Q6LChEREVksFhUiIiKyWCwqREREZLFYVIiIiMhisagQERGRxWJRISIiIovFokJEREQWi0WFiIiILBaLChFRI3r88cfh6uoqdwwiq8WiQmSlli9fDkmScPjwYbmjyOrxxx+HJEm33DQajdzxiKieVHIHICKqL7VajS+//LLafqVSKUMaImpILCpEZPVUKhUeeeQRuWMQUSPgrR8iG5eQkIDBgwdDq9XC1dUV/fr1w4EDB6qcU1ZWhjfffBPh4eHQaDTw8vJCr169sHXrVvM5GRkZmDhxIoKCgqBWqxEQEIDhw4fj0qVLt/27P/zwQ0iShMuXL1c7NnPmTDg6OuLGjRsAgHPnzmHUqFHw9/eHRqNBUFAQxo4di7y8vAYZh5u3ynbv3o2nn34aXl5e0Gq1eOyxx8wZ/uzzzz9HVFQU1Go1AgMDMXXqVOj1+mrnHTx4EPfddx88PDzg4uKC6Oho/Pvf/6523rVr1zBixAi4urrCx8cHL730EioqKhrkZyOyZbyiQmTDTp06hd69e0Or1eKVV16Bg4MDFi9ejHvuuQe7du1CbGwsAGDOnDmYO3cunnzySXTr1g0GgwGHDx/G0aNH0b9/fwDAqFGjcOrUKTz77LNo3rw5srKysHXrVly5cgXNmze/5d8/evRovPLKK1i1ahVefvnlKsdWrVqFAQMGwMPDA6WlpRg4cCCMRiOeffZZ+Pv749q1a9iwYQP0ej10Ot0df9acnJxq+xwdHaHVaqvsmzZtGtzd3TFnzhwkJydj4cKFuHz5Mnbu3AlJkszj8eabbyIuLg5Tpkwxn3fo0CH8/vvvcHBwAABs3boV999/PwICAjB9+nT4+/sjKSkJGzZswPTp081/Z0VFBQYOHIjY2Fh8+OGH2LZtGz766CO0bNkSU6ZMuePPRmTXBBFZpWXLlgkA4tChQ7c9Z8SIEcLR0VFcuHDBvC8tLU24ubmJPn36mPfFxMSIIUOG3PZzbty4IQCIDz74oNY5e/ToITp37lxlX3x8vAAgvv76ayGEEAkJCQKAWL16da0/f8KECQLALbeBAweaz7s5Xp07dxalpaXm/e+//74AINatWyeEECIrK0s4OjqKAQMGiIqKCvN5n376qQAgli5dKoQQory8XISFhYnQ0FBx48aNKplMJlO1fG+99VaVczp27FhtXIioOt76IbJRFRUV+PXXXzFixAi0aNHCvD8gIAAPP/ww9u7dC4PBAABwd3fHqVOncO7cuVt+lpOTExwdHbFz585b3ib5K2PGjMGRI0dw4cIF874ffvgBarUaw4cPBwDzFZMtW7agqKioVp8PABqNBlu3bq22zZs3r9q5kyZNMl8RAYApU6ZApVJh48aNAIBt27ahtLQUM2bMgELx//9EPvXUU9Bqtfjll18AVN5SS0lJwYwZM+Du7l7l77h5ZebPJk+eXOXr3r174+LFi7X+WYnsDYsKkY3Kzs5GUVERIiIiqh1r06YNTCYTUlNTAQBvvfUW9Ho9Wrdujfbt2+Pll1/G8ePHzeer1Wr885//xKZNm+Dn54c+ffrg/fffR0ZGxh1zPPTQQ1AoFPjhhx8AAEIIrF692vzcDACEhYXhhRdewJdffglvb28MHDgQn332WY2fT1EqlYiLi6u2dejQodq54eHhVb52dXVFQECA+Vmbm8/T/O+4OTo6okWLFubjN4tXu3bt7phPo9HAx8enyj4PD49alz4ie8SiQkTo06cPLly4gKVLl6Jdu3b48ssv0alTpypTfmfMmIGzZ89i7ty50Gg0mD17Ntq0aYOEhIS//OzAwED07t0bq1atAgAcOHAAV65cwZgxY6qc99FHH+H48eN47bXXUFxcjOeeew5RUVG4evVqw//ATYzTpInqjkWFyEb5+PjA2dkZycnJ1Y6dOXMGCoUCwcHB5n2enp6YOHEiVqxYgdTUVERHR2POnDlVvq9ly5Z48cUX8euvv+LkyZMoLS3FRx99dMcsY8aMwbFjx5CcnIwffvgBzs7OGDp0aLXz2rdvj3/84x/YvXs39uzZg2vXrmHRokW1/+H/wv/e3iooKEB6err5geDQ0FAAqDZupaWlSElJMR9v2bIlAODkyZMNmo+IqmJRIbJRSqUSAwYMwLp166pMIc7MzMT333+PXr16mW+9XL9+vcr3urq6olWrVjAajQCAoqIilJSUVDmnZcuWcHNzM5/zV0aNGgWlUokVK1Zg9erVuP/+++Hi4mI+bjAYUF5eXuV72rdvD4VCUaPPr40lS5agrKzM/PXChQtRXl6OwYMHAwDi4uLg6OiIjz/+GEII83lfffUV8vLyMGTIEABAp06dEBYWhgULFlSbtvzn7yOi+uH0ZCIrt3TpUmzevLna/unTp+Odd97B1q1b0atXLzzzzDNQqVRYvHgxjEYj3n//ffO5bdu2xT333IPOnTvD09MThw8fxo8//ohp06YBAM6ePYt+/fph9OjRaNu2LVQqFdasWYPMzEyMHTv2jhl9fX3Rt29fzJ8/H/n5+dVu++zYsQPTpk3DQw89hNatW6O8vBzffPMNlEolRo0adcfPLy8vx7fffnvLYyNHjqxSikpLS80/S3JyMj7//HP06tULw4YNA1B5JWrmzJl48803MWjQIAwbNsx8XteuXc0LyykUCixcuBBDhw5Fhw4dMHHiRAQEBODMmTM4deoUtmzZcsfcRFQDMs86IqI6ujnd9nZbamqqEEKIo0ePioEDBwpXV1fh7Ows+vbtK/bt21fls9555x3RrVs34e7uLpycnERkZKR49913zdN4c3JyxNSpU0VkZKRwcXEROp1OxMbGilWrVtU47xdffCEACDc3N1FcXFzl2MWLF8Xf/vY30bJlS6HRaISnp6fo27ev2LZt2x0/96+mJwMQKSkpVcZr165dYtKkScLDw0O4urqK8ePHi+vXr1f73E8//VRERkYKBwcH4efnJ6ZMmVJtGrIQQuzdu1f0799fuLm5CRcXFxEdHS0++eSTKvlcXFyqfd8bb7wh+E8w0Z1JQvAaJRHZvuXLl2PixIk4dOgQunTpInccIqohPqNCREREFotFhYiIiCwWiwoRERFZLD6jQkRERBaLV1SIiIjIYrGoEBERkcWy6gXfTCYT0tLS4Obmdsu3lRIREZHlEUIgPz8fgYGBVd5SfitWXVTS0tKqvKuEiIiIrEdqaiqCgoL+8hyrLipubm4AKn/Qm+8sISIiIstmMBgQHBxs/j3+V6y6qNy83aPVallUiIiIrExNHtvgw7RERERksVhUiIiIyGKxqBAREZHFYlEhIiIii8WiQkRERBaLRYWIiIgsFosKERERWSwWFSIiIrJYLCpERERksVhUiIiIyGKxqBAREZHFYlEhIiIii8WichvX9MW4mF0gdwwiIiK7xqJyC1/tTUHPeTvwr23n5I5CRERk11hUbqFzqAcAYHtSJopLK2ROQ0REZL9YVG4hJkiHYE8nFJVWYMeZLLnjEBER2S0WlVuQJAlD2gcCADYcT5M5DRERkf1iUbmN+6MDAAA7zmShwFgucxoiIiL7xKJyG1GBWoR5u8BYbsL2pEy54xAREdklFpXbkCTJfFXlv8fSZU5DRERkn1hU/sL90ZXPqew+m4284jKZ0xAREdkfFpW/EOHvhnBfV5RWmLD1NG//EBERNTUWlTu4eVWFs3+IiIiaHovKHdwfU/mcyt5zObhRWCpzGiIiIvvConIHLX1c0SZAi3KTwJZTGXLHISIisissKjVwc/bPhuOc/UNERNSUWFRqYOgfz6nsu5CD7HyjzGmIiIjsB4tKDYR4OSMm2B0mUflmZSIiImoaLCo19GzfVgCAZb+nID2vWOY0RERE9oFFpYb6tfFFt+aeMJab8K+tZ+WOQ0REZBdYVGpIkiS8OjgSAPDjkas4m5kvcyIiIiLbx6JSC51DPTAoyh8mAby/+YzccYiIiGwei0otvTwoAkqFhG1JWYhPyZU7DhERkU1jUamllj6uGNM1GAAwd1MShBAyJyIiIrJdLCp1MKNfOJwclEi4oudqtURERI2IRaUOfLUaPNk7DADw6k8n8CvLChERUaNgUamjp+9uiQ7B7sgrLsOkb45gzvpTMJZXyB2LiIjIprCo1JGrWoVVT/fAU39cWVm+7xJGLdyHlJxCmZMRERHZDklY8dOgBoMBOp0OeXl50Gq1suXYcSYTL646hhtFZXByUCK2hSc6h3igU6gHYoLd4apWyZaNiIjI0tTm9zeLSgPJyCvB9JUJOPg/U5YVUuVMoQh/N0T6uyHCX4tIfzcEeThBkiSZ0hIREcmHRUUmQgicuJaHo5dv4MgVPY5evoFr+lu/F8jTxRExQTp0CPZAhxB3dAxxh1bj0MSJiYiImh6LigXJNJTgdLoByRn5SM7Ix5mMfJzPykdZRdVhd1QqcHeED4bGBCKujS+cHXm7iIiIbBOLioUzllfgdJoBial6HEvV4+gVPa7kFpmPOzkoEdfWD6M6NUPvcB8oFbxFREREtoNFxQolZ+Tjv8fSsP5YWpXS0szdCQ91CcJDXYLRzN1JxoREREQNg0XFigkhcPxqHtYkXMOahGvIKy4DAEgS0C/SF38f3AatfF1lTklERFR3LCo2oqSsAltOZWBlfCr2X7wOAFApJDzRKwzP9gvntGciIrJKLCo26EJ2AeZuTMK2pCwAgJ9WjVlD2mJodACnORMRkVWpze9vrkxrJVr6uOLLCV2x9PEuCPF0RqbBiOdWJOCpr48gv6RM7nhERESNgkXFytwb6Ydfn++DF/q3hqNKgW1JmRj5OZfuJyIi28SiYoU0Dko81y8cq5/uAT+tGuezCjD8073YdTZb7mhEREQNStaiMmfOHEiSVGWLjIyUM5JViQl2x3+n9ULHEHcYSsoxcVk8vth9EVb82BEREVEVsl9RiYqKQnp6unnbu3ev3JGsiq9Wg5WTuuOhzkEwCeDdjUmYve4kKkwsK0REZP1kn9+qUqng7+8vdwyrplYp8f6D0YgM0OKdX07j2wNXkJNfigVjO0DjoJQ7HhERUZ3JfkXl3LlzCAwMRIsWLTB+/HhcuXLltucajUYYDIYqG1WSpMr1VT57uBMclQpsPpWBx76KR14RZwQREZH1krWoxMbGYvny5di8eTMWLlyIlJQU9O7dG/n5+bc8f+7cudDpdOYtODi4iRNbvvvaB+A/f+sGN7UK8Zdy8dDifUjPu/UbnImIiCydRS34ptfrERoaivnz5+OJJ56odtxoNMJoNJq/NhgMCA4OtosF32orKd2Ax5fFI9NgRDN3J6yc1B3Bns5yxyIiIrLeBd/c3d3RunVrnD9//pbH1Wo1tFptlY1urU2AFj9NuQth3i64pi/GmMX7cfk611ohIiLrYlFFpaCgABcuXEBAQIDcUWxCkIczVk7qjpY+LkjLK8GYxQe4MBwREVkVWYvKSy+9hF27duHSpUvYt28fRo4cCaVSiXHjxskZy6b4aTVYMak7wn1dkWEowZjF+3Ehu0DuWERERDUia1G5evUqxo0bh4iICIwePRpeXl44cOAAfHx85Ixlc3zdKstKhJ8bsvKNGLP4AC6yrBARkRWwqIdpa8ue3p7cEHILSzH+y4NISjcg2NMJP0/pCR83tdyxiIjIzljtw7TUuDxdHPHNE90Q6uWM1Nxi/G35IRQay+WORUREdFssKnbG21WN/0zsBk8XR5y4lodp3x9FeYVJ7lhERES3xKJih5p7u+CrCV2gcVDgt+RszF53ki8yJCIii8SiYqc6hnjgk3GdoJCAFfGp+GTHrdeuISIikhOLih3r39YPbw5vBwCYv/Usvj1wWeZEREREVbGo2LlHu4fi2XtbAQBmrzuJ/x5LkzkRERHR/2NRIbzQvzUe7R4KIYAXViVi19lsuSMREREBYFEhAJIk4c1hURgaE4iyCoHJ3xzBkcu5csciIiJiUaFKCoWEjx6Kwd2tfVBcVoGJyw4hKd0gdywiIrJzLCpk5qhSYNEjndE51AOGknI88uVBnM/KlzsWERHZMRYVqsLJUYmlj3dFu2ZaXC8sxcNfHOQbl4mISDYsKlSNzskB3/wtFpH+lS8xfPiLA0jNLZI7FhER2SEWFbolDxdHfPtkLFr6uCA9rwTjvjiANH2x3LGIiMjOsKjQbXm7qvH9U93R3MsZV28UY/yXB6EvKpU7FhER2REWFfpLfloNvn+qO5q5OyElpxDPrkjgSwyJiKjJsKjQHQW6O+HLCV3g5KDEnnM5mLfpjNyRiIjITrCoUI20CdBi/ugYAMCXe1Pw05GrMiciIiJ7wKJCNTa4fQCe++O9QDPXnEBiql7eQEREZPNYVKhWZsS1RlwbP5SWm/D0N4eRZSiROxIREdkwFhWqFYVCwr/GxCDc1xWZBiMmfXMEJWUVcsciIiIbxaJCteamccAXj3WBzskBial6zFpzEkIIuWMREZENYlGhOmnu7YLPx3eCUiHhp6NX8dXeFLkjERGRDWJRoTrr2cobs4e0AQC8tzEJO5OzZE5ERES2hkWF6mXCXc0xpkswTAJ4dkUCLmQXyB2JiIhsCIsK1YskSXhrRBS6hHogv6QcT/3nMJfZJyKiBsOiQvWmVimx8JHOCNRpcDGnEE9/cwTGcs4EIiKi+mNRoQbh46bG0old4aZW4WBKLl5efRwmE2cCERFR/bCoUIOJ9Ndi4SOdoVJIWH8sDR/+mix3JCIisnIsKtSgeoV7Y96oaADA5zsv4PuDV2RORERE1oxFhRrcg52DMCMuHAAwe91J/MZpy0REVEcsKtQopvcLx4Odg1BhEnj+h0Rk8p1ARERUBywq1CgkScJ7I9sjKlALfVEZXlp9jA/XEhFRrbGoUKNxVCnw77EdoFYpsOdcDv6z/5LckYiIyMqwqFCjauXrhll/LLM/d9MZnM3MlzkRERFZExYVanSPdg/F3a19UFpuwvSViVwMjoiIaoxFhRqdJEn44KFoeLo4IindgPm/npU7EhERWQkWFWoSvm4azH2gPQBgyZ6L2H/husyJiIjIGrCoUJMZGOWPMV2CIQTw0upjyC8pkzsSERFZOBYValKzh7ZFkIcTrumL8c6GJLnjEBGRhWNRoSblqlbhw4diIEnAD4dTsT0pU+5IRERkwVhUqMl1b+GFJ3qGAQBe/ekEcgtLZU5ERESWikWFZPHSwAi08nVFToERs9eehBBctZaIiKpjUSFZaByUmD86BkqFhF9OpGP9sTS5IxERkQViUSHZRAe549l7WwEA5qw/xVtARERUDYsKyWpq31aI9HfDjaIyvPsLZwEREVFVLCokKwelAu890B6SBPx09Cr2nc+ROxIREVkQFhWSXacQDzwSGwoAmLX2JErK+C4gIiKqxKJCFuHlQRHwdVMjJacQn++8IHccIiKyECwqZBG0GgfMGRYFAFi48zzOZxXInIiIiCwBiwpZjMHt/HFvpC/KKgReW3MCJhPXViEisncsKmQxJEnCW8Oj4OSgRHxKLn44nCp3JCIikpnFFJV58+ZBkiTMmDFD7igkoyAPZ7w4oDUA4L1fkpCRVyJzIiIikpNFFJVDhw5h8eLFiI6OljsKWYCJPcPQIdgd+cZyzFpzgsvrExHZMdmLSkFBAcaPH48vvvgCHh4ecschC6BUSHj/wWg4KhXYfiaLy+sTEdkx2YvK1KlTMWTIEMTFxd3xXKPRCIPBUGUj29Taz63K8vo5BUaZExERkRxkLSorV67E0aNHMXfu3BqdP3fuXOh0OvMWHBzcyAlJTpPvaYk2AVrcKCrDnPWn5I5DREQykK2opKamYvr06fjuu++g0Whq9D0zZ85EXl6eeUtN5awQW+agVOCDB6OhVEjYcDwdW05lyB2JiIiamCRkelJx7dq1GDlyJJRKpXlfRUUFJEmCQqGA0WiscuxWDAYDdDod8vLyoNVqGzsyyeT9zWfw+c4L8HVT47eX7oGLWiV3JCIiqofa/P6W7YpKv379cOLECSQmJpq3Ll26YPz48UhMTLxjSSH78Vy/cIR6OSMr34jFu7i8PhGRPZGtqLi5uaFdu3ZVNhcXF3h5eaFdu3ZyxSILpHFQYubgSADAkj0XkaYvljkRERE1Fdln/RDVxMAof3QL80RJmQkfbEmWOw4RETURiyoqO3fuxIIFC+SOQRZIkiTMHtIWkgSsSbiGxFS93JGIiKgJWFRRIfor7YN0eKBjEADgnQ2nuWItEZEdYFEhq/LKoAg4OShx+PIN/HIiXe44RETUyFhUyKr4aTWYfHdLAMC8TWdQUlYhcyIiImpMLCpkdZ7qEwZ/rQZXbxRj+b5LcschIqJGxKJCVsfZUYWXBkYAABbuvIC84jKZExERUWNhUSGrNLJjM7T2c0VecRkXgSMismEsKmSVlAoJLw+sXARu6e8pyDKUyJyIiIgaA4sKWa24Nr7oFOKOkjITPt5xTu44RETUCFhUyGpJkoRXB1VeVVkZn4pLOYUyJyIioobGokJWLbaFF+6J8EG5SWD+1rNyxyEiogbGokJW75U/nlVZfywNp9LyZE5DREQNiUWFrF7bQC2GdwgEALy/mS8sJCKyJSwqZBNe6N8aKoWEXWezsedcttxxiIiogbCokE0I9XLBYz2aAwDe2ZCE8gqTvIGIiKhBsKiQzZjeLxzuzg5IzszHykOpcschIqIGwKJCNkPn7IDn41oDAOZvPcul9YmIbACLCtmUh2ND0MrXFbmFpfjst/NyxyEionpiUSGb4qBUYNaQNgCAZb+ncBE4IiIrx6JCNqdvhC/ubu2DsgqB9zYmyR2HiIjqgUWFbNI/hrSBUiHh19OZ2HchR+44RERURywqZJPC/dwwPjYEQOV05QqTkDkRERHVBYsK2awZca3hplHhdLoBPx25KnccIiKqAxYVslmeLo6Y3i8cAPDBr8koNJbLnIiIiGqLRYVs2qM9QhHq5YzsfCMW7bogdxwiIqolFhWyaWqVEjMHV05XXrL7ItL0xTInIiKi2mBRIZs3MMoPsWGeMJab8P7mM3LHISKiWmBRIZsnSRJm398WkgSsTUxDYqpe7khERFRDLCpkF9o10+GBjkEAgLc3nIYQnK5MRGQNWFTIbrw8MAJODkocuXwDm09myB2HiIhqgEWF7Ia/ToOn+rQAAPxz8xmUlptkTkRERHfCokJ2ZVKfFvB2dcSl60VYEX9F7jhERHQHLCpkV1zVKsyIaw0A+Pf2czCUlMmciIiI/gqLCtmdsV2D0cLHBbmFpVi0k4vAERFZMhYVsjsqpQJ/HxQJAPhqbwrS87gIHBGRpWJRIbvUv60fujWvXATuo1/Pyh2HiIhug0WF7JIkSZh5X+VVlZ+OXkVSukHmREREdCssKmS3OoZ4YEh0AIQA3tuYxEXgiIgsEIsK2bVXB0bCUanAnnM5+C05S+44RET0P1hUyK6FeDljYq/mAIB3NiShrIKLwBERWRIWFbJ70/q2grerIy7mFOKb/ZfljkNERH/CokJ2z03jgBcHRAAAFmw7ixuFpTInIiKim1hUiACM7hKMSH83GErKsWAbpysTEVkKFhUiAEqFhNeHtgUAfHvwCs5l5suciIiIABYVIrO7WnpjQFs/VJgE3vklSe44REQEFhWiKmYNaQMHpYRdZ7Ox62y23HGIiOweiwrRn4R6ueCxHs0BAB9sOQOTiYvAERHJiUWF6H88c09LuDgqcfKaAZtPZcgdh4jIrtWpqKSmpuLq1avmr+Pj4zFjxgwsWbKkwYIRycXLVY0ne7cAAHz4azLKuQgcEZFs6lRUHn74Yfz2228AgIyMDPTv3x/x8fGYNWsW3nrrrQYNSCSHJ3uHwcPZARezC/Hz0WtyxyEislt1KionT55Et27dAACrVq1Cu3btsG/fPnz33XdYvnx5Q+YjkoWbxgHP3NMKQOUicCVlFTInIiKyT3UqKmVlZVCr1QCAbdu2YdiwYQCAyMhIpKenN1w6Ihk92iMU/loN0vJK8N3BK3LHISKyS3UqKlFRUVi0aBH27NmDrVu3YtCgQQCAtLQ0eHl51fhzFi5ciOjoaGi1Wmi1WvTo0QObNm2qSySiBqdxUGJ6XDgA4LPfzqPAWC5zIiIi+1OnovLPf/4Tixcvxj333INx48YhJiYGALB+/XrzLaGaCAoKwrx583DkyBEcPnwY9957L4YPH45Tp07VJRZRg3uwcxCaezkjt7AUX+1JkTsOEZHdkYQQdVoooqKiAgaDAR4eHuZ9ly5dgrOzM3x9fescyNPTEx988AGeeOKJO55rMBig0+mQl5cHrVZb57+T6K+sP5aG51YkQK1SYNP03mjh4yp3JCIiq1ab3991uqJSXFwMo9FoLimXL1/GggULkJycXOeSUlFRgZUrV6KwsBA9evSo02cQNYah0QHo1cobxnITXv7xOCq4CBwRUZOpU1EZPnw4vv76awCAXq9HbGwsPvroI4wYMQILFy6s1WedOHECrq6uUKvVmDx5MtasWYO2bdve8lyj0QiDwVBlI2pskiRh3qj2cFWrcOTyDSz7nbeAiIiaSp2KytGjR9G7d28AwI8//gg/Pz9cvnwZX3/9NT7++ONafVZERAQSExNx8OBBTJkyBRMmTMDp06dvee7cuXOh0+nMW3BwcF3iE9VakIczXruvDQDggy3JuJBdIHMiIiL7UKdnVJydnXHmzBmEhIRg9OjRiIqKwhtvvIHU1FRERESgqKiozoHi4uLQsmVLLF68uNoxo9EIo9Fo/tpgMCA4OJjPqFCTEELgsaXx2HMuB51C3LF68l1QKiS5YxERWZ1Gf0alVatWWLt2LVJTU7FlyxYMGDAAAJCVlVXvwmAymaqUkT9Tq9Xmqcw3N6KmUnkLKBquahWOXtFj6V7eAiIiamx1Kiqvv/46XnrpJTRv3hzdunUzP/z666+/omPHjjX+nJkzZ2L37t24dOkSTpw4gZkzZ2Lnzp0YP358XWIRNbpm7k6YNaTyFtCHvybjIm8BERE1qjpPT87IyEB6ejpiYmKgUFT2nfj4eGi1WkRGRtboM5544gls374d6enp0Ol0iI6Oxquvvor+/fvX6Ps5PZnk8OdbQH1a++A/E7tCkngLiIiopmrz+7vOReWmm29RDgoKqs/H1AmLCsklJacQA/61C2UVAl8+1gVxbf3kjkREZDUa/RkVk8mEt956CzqdDqGhoQgNDYW7uzvefvttmEymOoUmsiZh3i54olcLAMBbG07zpYVERI2kTkVl1qxZ+PTTTzFv3jwkJCQgISEB7733Hj755BPMnj27oTMSWaRp97aCr5saV3KL8BUfrCUiahR1uvUTGBiIRYsWmd+afNO6devwzDPP4Nq1aw0W8K/w1g/JbU3CVTz/wzE4OSix46W7EaBzkjsSEZHFa/RbP7m5ubd8YDYyMhK5ubl1+UgiqzSiQzN0CnFHcVkF5m06I3ccIiKbU6eiEhMTg08//bTa/k8//RTR0dH1DkVkLSRJwpvD2kGSgHWJaTh0iUWdiKghqeryTe+//z6GDBmCbdu2mddQ2b9/P1JTU7Fx48YGDUhk6doH6TC2azBWxKdi9tqTWDetJ9QqpdyxiIhsQp2uqNx99904e/YsRo4cCb1eD71ejwceeACnTp3CN99809AZiSzeSwMi4OniiDMZ+bwFRETUgOq9jsqfHTt2DJ06dUJFRdNM1eTDtGRJdpzJxN+WHwYAfPFYF/Tn2ipERLfU6A/TElF190b64cleYQCAl388hjR9scyJiIisH4sKUQN6ZVAkooN00BeVYcbKRJRXcAFEIqL6YFEhakCOKgU+GdcRrmoV4i/l4uMd5+WORERk1Wo16+eBBx74y+N6vb4+WYhsQqiXC957oD2eW5GAT3acQ8+WXoht4SV3LCIiq1SroqLT6e54/LHHHqtXICJbMCwmEHvPZWPV4at4fd0p/PJcL6iUvIBJRFRbDTrrp6lx1g9ZMn1RKe75cCf0RWV4e3gUHu3RXO5IREQWgbN+iCyAu7MjXuzfGgDw0daz0BeVypyIiMj6sKgQNaJx3UIQ4ecGfVEZFmw7J3ccIiKrw6JC1IhUSgVeH9oWAPDNgcs4l5kvcyIiIuvCokLUyHq28saAtn6oMAm8teE0rPixMCKiJseiQtQEZg1pA0elAnvO5WB7UpbccYiIrAaLClETCPVywRO9K5fXf/uX0ygubZr3YRERWTsWFaImMrVvK/hp1bh8vQjvb+EblomIaoJFhaiJuKpVmDcqGgCw7PdL2HchR+ZERESWj0WFqAn1jfDFuG4hAICXVx9HfkmZzImIiCwbiwpRE5s1pA2CPZ1wTV+MtzecljsOEZFFY1EhamKuahU+eqgDJAlYdfgqtp3OlDsSEZHFYlEhkkG3ME882atyFtDffz6B3EIur09EdCssKkQyeXFABMJ9XZFTYMQ/1p7gQnBERLfAokIkE42DEvNHd4BKIWHjiQysS0yTOxIRkcVhUSGSUfsgHZ7rFw4AmL3uJNLzimVORERkWVhUiGT2zD0tERPsjvyScry8+jhMJt4CIiK6iUWFSGYqpQLzR8dA46DA3vM5+ObAZbkjERFZDBYVIgvQ0scVMwe3AQDM3ZSEC9kFMiciIrIMLCpEFuLR7qHoHe6NkjITXlh1DGUVJrkjERHJjkWFyEIoFBLefzAaWo0Kx1L1+GTHebkjERHJjkWFyIIE6Jzwzsj2AIBPd5zDkcu5MiciIpIXiwqRhRkWE4iRHZvBJIAZPyTyxYVEZNdYVIgs0JvDo9DM3QmpucV4Y/0pueMQEcmGRYXIAmk1DlgwtgMUEvDz0WvYcJyr1hKRfWJRIbJQXZt74pl7WgEAXvv5BNL0XLWWiOwPiwqRBZseF46YIB0MJeWYsTIR5ZyyTER2hkWFyII5KBVYMLYjXByViL+Ui/lbz8odiYioSbGoEFm4MG8XzBsVDQD4fOcF/JacJXMiIqKmw6JCZAWGxgTi0e6hAIDnf0jk8ypEZDdYVIisxD/ub4P2zXTQF5Vh2vdHucQ+EdkFFhUiK6FWKfHZw53gplHh6BU93t98Ru5IRESNjkWFyIqEeDnjgwdjAABf7EnBr6cyZE5ERNS4WFSIrMygdv74W88wAMBLq48hNbdI5kRERI2HRYXICv19cCRigt1hKCnHtBUJKC3n8ypEZJtYVIiskKNKgU/HdYRWo8KxVD3+yedViMhGsagQWalgT2d8NLoDAOCrvXxehYhsE4sKkRXr39YPT/bi8ypEZLtYVIis3KuDI9Ex5P+fV+H6KkRkS2QtKnPnzkXXrl3h5uYGX19fjBgxAsnJyXJGIrI6DkoFPvnT8yr/3nZO7khERA1G1qKya9cuTJ06FQcOHMDWrVtRVlaGAQMGoLCwUM5YRFYnyMMZcx+ofB/QZzvP48DF6zInIiJqGJIQQsgd4qbs7Gz4+vpi165d6NOnzx3PNxgM0Ol0yMvLg1arbYKERJbt5dXHsPrIVQToNNg8vQ90zg5yRyIiqqY2v78t6hmVvLw8AICnp+ctjxuNRhgMhiobEf2/OcOiEObtgvS8Ery25gQs6P8PISKqE4spKiaTCTNmzEDPnj3Rrl27W54zd+5c6HQ68xYcHNzEKYksm4tahQVjOkClkPDLiXSsPnJV7khERPViMUVl6tSpOHnyJFauXHnbc2bOnIm8vDzzlpqa2oQJiaxDTLA7XhjQGgAwZ/0pnMnglUcisl4WUVSmTZuGDRs24LfffkNQUNBtz1Or1dBqtVU2Iqru6T4t0aOFF4pKK/DwFwdZVojIaslaVIQQmDZtGtasWYMdO3YgLCxMzjhENkOpkLDokc6IDtIht7AU45YcQFI6ywoRWR9Zi8rUqVPx7bff4vvvv4ebmxsyMjKQkZGB4uJiOWMR2QSdswO+eSIWMUE63Cgqw8NfHMDpNJYVIrIusk5PliTplvuXLVuGxx9//I7fz+nJRHeWV1yGx5bG41iqHu7ODvjuyVhEBerkjkVEdsxqpicLIW651aSkEFHN6Jwc8M0T3dAh2B36ojI8+lU8Ll/noopEZB0s4mFaImpcWo0Dvn6iG9o10yK3sBQTlx2CvqhU7lhERHfEokJkJ7QaByyd0BXN3J1wMacQk74+AmN5hdyxiIj+EosKkR3x1Wqw9PGucFOrEH8pF6/8eJyr1xKRRWNRIbIzEf5uWPhIZ6gUEtYlpmH+1rNyRyIiui0WFSI71CvcG++NbA8A+GTHeayMvyJzIiKiW2NRIbJTo7sG49l7WwEAXltzAptPpsuciIioOhYVIjv2Qv/WGNs1GCYBPLciEfsu5MgdiYioChYVIjsmSRLeGdEOg6L8UVphwqSvj+DktTy5YxERmbGoENk5lVKBBWM7oEcLLxQYyzFhaTwuZhfIHYuICACLChEB0DgoseSxzmjXTIvrhaV49Kt4pOYWyR2LiIhFhYgquWkcsHxiN7TwccE1fTHGLjmAK9dZVohIXiwqRGTm7arGiqe6/6ms7Od7gYhIViwqRFSFn1aDlU91R0sfF6TllWDskgMsK0QkGxYVIqrGV6vBiknd0crXFel5JRiz+ABSclhWiKjpsagQ0S35ummw4qnuCPd1RYahBA8u3IcTVzl1mYiaFosKEd2Wj5saKyZ1R1Rg5WygsUv2Y+85LgpHRE2HRYWI/pK3qxorJ3XHXS29UFhagYnL47H+WJrcsYjITrCoENEduWkcsGxiVwyJDkBZhcBzKxKw7PcUuWMRkR1gUSGiGlGrlPhkbEdM6BEKAHjzv6fx1V6WFSJqXCwqRFRjCoWEOcOi8Nwfb11+e8NpfHfwssypiMiWsagQUa1IkoTn+7fG031aAAD+sfYkfjpyVeZURGSrWFSIqNYkScLfB0diQo9QCAG8/OMxbDjOB2yJqOGxqBBRnUiShDeGRmFs12CYBDBjZSI2nkiXOxYR2RgWFSKqM4VCwrsj22NEh0CUmwSe+e4o/r3tHEwmIXc0IrIRLCpEVC9KhYQPH4rB43c1BwD8a9tZTP3+KAqN5fIGIyKbwKJCRPWmUiowZ1gU3h8VDQelhE0nMzBq4T6k5hbJHY2IrByLChE1mNFdg7FyUnd4u6pxJiMfwz7di/iUXLljEZEVY1EhogbVOdQT66f1RPtmOtwoKsP4Lw9g9eFUuWMRkZViUSGiBhfo7oRVT/fAfe39UVYh8PKPx/HPzWf4kC0R1RqLChE1CidHJT4d1wnP/rGK7cKdFzDluyMoKuVDtkRUcywqRNRoFAoJLw6IwL/GxMBRqcCWU5l4cOF+XL3Bh2yJqGZYVIio0Y3sGITvn4qFl4sjTqcbMOzT33Hw4nW5YxGRFWBRIaIm0aW5J9Y/2wtRgVrkFpZi/JcH8c0BvtCQiP4aiwoRNZlm7k74cfJdGBpTuZLt7LUnMfPnEygtN8kdjYgsFIsKETUpJ0clPh7bAa8OioQkASvir+DhLw4gK79E7mhEZIFYVIioyUmShCn3tMTSCV3hplHh8OUbGPbJ70hM1csdjYgsDIsKEcmmb6Qv1k3tiVa+rsgwlGD0ov1YxcXhiOhPWFSISFYtfFyx5pm7MKCtH0orTHjlx+OY+fNxGErK5I5GRBaARYWIZOemccCiRzrj+bjWAIAV8ano99EurD+WBiG4mi2RPWNRISKLoFBImB4Xju+fikULbxdk5xvx3IoEPLY0HpdyCuWOR0QyYVEhIotyV0tvbJrRG8/HtYajSoE953IwYMFufLH7It8VRGSHWFSIyOKoVUpMjwvHlhl90KuVN0rLTXh3YxIe/vIAl98nsjMsKkRkscK8XfDNE93w3sj2cHZU4sDFXAxesAc/HbnKZ1eI7ASLChFZNEmS8HBsCDZN741OIe7IN5bjxdXH8Mx3R5FXxJlBRLaORYWIrEKolwtWPd0DLw+MgEohYdPJDAz5ZA8SrtyQOxoRNSIWFSKyGiqlAlP7tsKaZ3oixNMZV28U46FF+/HF7ou8FURko1hUiMjqtA/SYcNzvTAkOgDlJoF3Nybhyf8cxvUCo9zRiKiBsagQkVXSahzw6biOeHdkOziqFNh+Jgtx83fxQVsiG8OiQkRWS5IkjI8NxdpneiLS3w03isrw4upjeGxpPK5c5zRmIlvAokJEVq9toBb/fbYXXhkUAbV5kbhdWLTrAsoqTHLHI6J6YFEhIpvgoFTgmXtaYcuMPrirpRdKykyYt+kMhn6yF0cuc2YQkbWStajs3r0bQ4cORWBgICRJwtq1a+WMQ0Q2oLm3C757MhbvPxgND2cHnMnIx4OL9uG1NSe47gqRFZK1qBQWFiImJgafffaZnDGIyMZIkoTRXYKx/cV78GDnIAgBfH/wCvrN34m1Cdf4sC2RFZGEhfwvVpIkrFmzBiNGjKjx9xgMBuh0OuTl5UGr1TZeOCKyagcuXsesNSdwIbvyLczdW3ji7eHtEO7nJnMyIvtUm9/fVvWMitFohMFgqLIREd1J9xZe2Di9N14eGAGNg6LynUH/3oO5G5NQaCyXOx4R/QWrKipz586FTqczb8HBwXJHIiIroVYpMbVvK2x9/m4MaOuHcpPA4t0Xce9HO/GffZdQUlYhd0QiugWruvVjNBphNP7/ypMGgwHBwcG89UNEtbbjTCbmrD+NK7mV6634uqnx9N0t8XC3EDg5KmVOR2TbanPrR9VEmRqEWq2GWq2WOwYR2YB7I/3Qs5U3Vh++ioU7L+CavhhvbziNhTvPY/LdLfFoj1CoVSwsRHKzqls/REQNSa1S4pHuofjtpXsw74H2CPZ0Qk5BKd75JQn9PtqFdYnXYDJZxEVnIrsla1EpKChAYmIiEhMTAQApKSlITEzElStX5IxFRHbGUaXA2G4h2PFiZWHx06px9UYxpq9MxLDP9mLvuRxOaSaSiazPqOzcuRN9+/attn/ChAlYvnz5Hb+f05OJqDEUl1Zg6e8pWLjzAgr+mBXUNkCLx3qEYniHZnyGhaieavP722Iepq0LFhUiakzXC4z4ZMd5rIi/AmN55TuDtBoVRncJxoS7miPY01nmhETWiUWFiKgB6YtKsepwKr49cMU8S0ilkPBQlyBMuzcczdydZE5IZF1YVIiIGoHJJLDrbDaW/p6CPedyAAAOSglju4Zgat9W8NdpZE5IZB1YVIiIGtnhS7mYv/Us9l24DqDygdwxXYIxqU8L3hIiugMWFSKiJrLvQg7m/3oWhy/fAAAoFRKGdwjElLtb8l1CRLfBokJE1ISEENh/8To+/+0C9p7PMe8f0NYPT/VpgS6hHpAkScaERJaFRYWISCbHUvX4fOd5bDmVad4XE+yOJ3uFYXA7f6iUXGeTiEWFiEhm57Py8dXeFPx09BpK/5ja3MzdCQ/HhmBUpyA+eEt2jUWFiMhC5BQY8e2By/hm/2VcLywFACgk4J4IX4zpGox7I33hwKssZGdYVIiILExJWQU2HE/HqkOpiL+Ua97v7eqI+6MD8UCnZmjfTMdnWcgusKgQEVmwC9kFWHU4FT8duYacAqN5f0sfF4zs2AwjOjZDkAenOJPtYlEhIrICZRUm7D2Xg58TruHXUxnmZfoBoFuYJx7o2AyD2wdA5+QgY0qihseiQkRkZfJLyrD5ZAbWJFzD/ovXcfNfZkeVAv3b+GFst2D0bOkNhYK3hsj6sagQEVmxNH0x1iWmYU3CVZzNLDDvD/Z0wtiuIXiocxB8tZw1RNaLRYWIyAYIIXAqzYDVh1Pxc8I15JeUA6hc/bZvhA9GdgxCvza+0DgoZU5KVDssKkRENqa4tAK/nEjHivgrOPLHcv0A4KZR4b52ARjRsRm6NvfggnJkFVhUiIhs2LnMfPyccA3rEq4hLa/EvN9NrULPVt64O8IHfVr7oJm7k4wpiW6PRYWIyA6YTAIHU3KxJuEqfj2dCX1RWZXj4b6uuDfSF30jfdE51IMLy5HFYFEhIrIzFSaBE9fysCs5G7vPZSPhyg2Y/vSvu5tGhT6tfTAoyh99I33hqlbJF5bsHosKEZGdyysqw+5z2fjtTBZ2ns1G7h/L9wOVU57vbu2D+9r7495IP67TQk2ORYWIiMwqTALHruqx9XQmNp1Ix6XrReZjKoWEbmGe6NfGD3FtfBHq5SJjUrIXLCpERHRLQgicycjHphPp2HgyA+ezCqocb+njgg7BHmgT4IY2AVpE+rvBy1UtU1qyVSwqRERUI5dyCrEtKRPbk7Jw6FIuyk3VfyX4uqkRHaRD+2bulX8G6eDN8kL1wKJCRES1lldchoMXryMpPR9J6QacyTDgcm4RbvVbooW3C2JbeKF7C0/0aOHFlXKpVlhUiIioQRQay5GUbsDxq3k4cS0Px6/qcTGnsFp5aeHjgt6tvNEr3AfdW3jCTcMHdOn2WFSIiKjR5BWX4fClXOy/cB0HUq7jVJqhSnFRKSR0CHZH5+YeaBugRVSgFmHerlDyhYr0BxYVIiJqMnlFZdh/8Tr2ns/GnnM5uPynWUU3aRwUiPTXokOwOzqGuKNDsDtCPJ0hSSwv9ohFhYiIZJOaW4Tfz+fgZFoeTqcZcCYjH0WlFdXO83RxRLtmOkT6uyHS3w0R/m5o5esKtYovWbR1LCpERGQxTCaBy7lFOH5Vj8RUPRKu6HE6zYDSClO1c5UKCcEeTmjh44oW3i5o4eOKcD9XtPZz48J0NoRFhYiILJqxvAKn0wxISs9HcoYBSRn5SM7IR15x2W2/J1CnQYS/GyL8tWjfTIfoIB2CPJx4+8gKsagQEZHVEUIg02DExewCXMgprPwzuxDnMvOR/qe3RP+Zp4sj2jfToU2AFsGeTgj2cEaQhxMC3Z2gceAtJEvFokJERDYlr7gMZzPzcSajco2XE1fzcCbDgLKKW/8KkySgmbsTWvm6opWPK1r5Vt5CCvdzg5ZTp2XHokJERDavpKwCZzLycfyqHheyCnD1RjFSbxQhNbcYxWXVH9696eYtpNb+bgjxdEaAToMAnRMCdBronBx4K6kJsKgQEZHdEkLgemEpLmQV4Hx2Ac5nVW7nMguQYbj1LaSbVAoJSoUEhSRBIQEKSUKQpzO6hHqgS3MPdG3uiUB3pyb6SWwXiwoREdEt3LyFlJyRj3OZ+bimL0Z6XgnS80qQW1hao88I0GnQyrdyVlLYHzOTmnu5IMBdAwelopF/AtvAokJERFRLJWUVuFFUCpOonFItBFBmMuFsRj4OXbqBw5dzcSrNgIpbvLgRqJxaHaDTIMTTGSGezmh+s8h4uyDEy5nrw/wJiwoREVEjKDSW43S6ASnZhbj4x8ykizmFuJJbhNLy6uvC3KSQAB83NXRODnB3coTWyQHuzg7w12oQ4K5BoLsTAnVOCHDXwE2tsvnnZGrz+1vVRJmIiIisnotaha7NPdG1uWeV/SaTQHaBEVdyi5CaW4RL14twKacQKX9sBcZyZBqMyDQY7/h3ODsq4afVwE+rhr9WA3fnymKj1aj++NMBni6O8HRxgLuzI9ydHKCy4VtOLCpERET1pFBIf5QLTbUSI0RlickyGKEvKkNeceV2o6gUGXklSNMXI+2PP/OKy1BUWmEuODXl7uwALxdHeLmo4eXqCC9XRwTonCqv2Og08Ndp4OWqhptaBYWVvRySRYWIiKgRSZIEXzcNfN00dzy3qPTmlZcS85ZXXAZDcTkMJWUwFJdBX1wGfVEZcgtLzSv56osq913I/utyo5AAN03lbSedU+UVGU/nP/50cYSHswN0f1ylcXeuvE3l6eoIV7V8dYFFhYiIyEI4O6oQ5q1CmLdLjc4vrzAhr7iytOQUlOJ6oRG5haXIzjciPa8EGXklSM8rRkZeCQpLK2ASMF/RqamBUX5Y/GiXuv5I9caiQkREZKVUSgW8XNXwclUj3O+vzy0pq4Ch+P9vPemLKm8/6YvKkFtUihuFlf+tLy4136LSF5XB3cmxaX6Y22BRISIisgMaByU0Dkr4au98C+rPbjcdu6nY7mPCREREVG9KmR++ZVEhIiIii8WiQkRERBaLRYWIiIgsFosKERERWSwWFSIiIrJYLCpERERksVhUiIiIyGKxqBAREZHFsoii8tlnn6F58+bQaDSIjY1FfHy83JGIiIjIAsheVH744Qe88MILeOONN3D06FHExMRg4MCByMrKkjsaERERyUz2ojJ//nw89dRTmDhxItq2bYtFixbB2dkZS5culTsaERERyUzWolJaWoojR44gLi7OvE+hUCAuLg779++vdr7RaITBYKiyERERke2Stajk5OSgoqICfn5V303t5+eHjIyMaufPnTsXOp3OvAUHBzdVVCIiIpKBSu4AtTFz5ky88MIL5q/z8vIQEhLCKytERERW5ObvbSHEHc+Vtah4e3tDqVQiMzOzyv7MzEz4+/tXO1+tVkOtVpu/vvmD8soKERGR9cnPz4dOp/vLc2QtKo6OjujcuTO2b9+OESNGAABMJhO2b9+OadOm3fH7AwMDkZqaCjc3N0iS1KDZDAYDgoODkZqaCq1W26CfTVVxrJsOx7rpcKybDse66TTUWAshkJ+fj8DAwDueK/utnxdeeAETJkxAly5d0K1bNyxYsACFhYWYOHHiHb9XoVAgKCioUfNptVr+H34T4Vg3HY510+FYNx2OddNpiLG+05WUm2QvKmPGjEF2djZef/11ZGRkoEOHDti8eXO1B2yJiIjI/sheVABg2rRpNbrVQ0RERPZF9gXfLJVarcYbb7xR5eFdahwc66bDsW46HOumw7FuOnKMtSRqMjeIiIiISAa8okJEREQWi0WFiIiILBaLChEREVksFhUiIiKyWCwqt/DZZ5+hefPm0Gg0iI2NRXx8vNyRrN7cuXPRtWtXuLm5wdfXFyNGjEBycnKVc0pKSjB16lR4eXnB1dUVo0aNqvZ6Baq9efPmQZIkzJgxw7yPY91wrl27hkceeQReXl5wcnJC+/btcfjwYfNxIQRef/11BAQEwMnJCXFxcTh37pyMia1TRUUFZs+ejbCwMDg5OaFly5Z4++23q7wrhmNdd7t378bQoUMRGBgISZKwdu3aKsdrMra5ubkYP348tFot3N3d8cQTT6CgoKD+4QRVsXLlSuHo6CiWLl0qTp06JZ566inh7u4uMjMz5Y5m1QYOHCiWLVsmTp48KRITE8V9990nQkJCREFBgfmcyZMni+DgYLF9+3Zx+PBh0b17d3HXXXfJmNr6xcfHi+bNm4vo6Ggxffp0836OdcPIzc0VoaGh4vHHHxcHDx4UFy9eFFu2bBHnz583nzNv3jyh0+nE2rVrxbFjx8SwYcNEWFiYKC4uljG59Xn33XeFl5eX2LBhg0hJSRGrV68Wrq6u4t///rf5HI513W3cuFHMmjVL/PzzzwKAWLNmTZXjNRnbQYMGiZiYGHHgwAGxZ88e0apVKzFu3Lh6Z2NR+R/dunUTU6dONX9dUVEhAgMDxdy5c2VMZXuysrIEALFr1y4hhBB6vV44ODiI1atXm89JSkoSAMT+/fvlimnV8vPzRXh4uNi6dau4++67zUWFY91wXn31VdGrV6/bHjeZTMLf31988MEH5n16vV6o1WqxYsWKpohoM4YMGSL+9re/Vdn3wAMPiPHjxwshONYN6X+LSk3G9vTp0wKAOHTokPmcTZs2CUmSxLVr1+qVh7d+/qS0tBRHjhxBXFyceZ9CoUBcXBz2798vYzLbk5eXBwDw9PQEABw5cgRlZWVVxj4yMhIhISEc+zqaOnUqhgwZUmVMAY51Q1q/fj26dOmChx56CL6+vujYsSO++OIL8/GUlBRkZGRUGWudTofY2FiOdS3ddddd2L59O86ePQsAOHbsGPbu3YvBgwcD4Fg3ppqM7f79++Hu7o4uXbqYz4mLi4NCocDBgwfr9fdbxBL6liInJwcVFRXV3jPk5+eHM2fOyJTK9phMJsyYMQM9e/ZEu3btAAAZGRlwdHSEu7t7lXP9/PyQkZEhQ0rrtnLlShw9ehSHDh2qdoxj3XAuXryIhQsX4oUXXsBrr72GQ4cO4bnnnoOjoyMmTJhgHs9b/ZvCsa6dv//97zAYDIiMjIRSqURFRQXeffddjB8/HgA41o2oJmObkZEBX1/fKsdVKhU8PT3rPf4sKtTkpk6dipMnT2Lv3r1yR7FJqampmD59OrZu3QqNRiN3HJtmMpnQpUsXvPfeewCAjh074uTJk1i0aBEmTJggczrbsmrVKnz33Xf4/vvvERUVhcTERMyYMQOBgYEcaxvHWz9/4u3tDaVSWW32Q2ZmJvz9/WVKZVumTZuGDRs24LfffkNQUJB5v7+/P0pLS6HX66ucz7GvvSNHjiArKwudOnWCSqWCSqXCrl278PHHH0OlUsHPz49j3UACAgLQtm3bKvvatGmDK1euAIB5PPlvSv29/PLL+Pvf/46xY8eiffv2ePTRR/H8889j7ty5ADjWjakmY+vv74+srKwqx8vLy5Gbm1vv8WdR+RNHR0d07twZ27dvN+8zmUzYvn07evToIWMy6yeEwLRp07BmzRrs2LEDYWFhVY537twZDg4OVcY+OTkZV65c4djXUr9+/XDixAkkJiaaty5dumD8+PHm/+ZYN4yePXtWm2Z/9uxZhIaGAgDCwsLg7+9fZawNBgMOHjzIsa6loqIiKBRVf2UplUqYTCYAHOvGVJOx7dGjB/R6PY4cOWI+Z8eOHTCZTIiNja1fgHo9imuDVq5cKdRqtVi+fLk4ffq0mDRpknB3dxcZGRlyR7NqU6ZMETqdTuzcuVOkp6ebt6KiIvM5kydPFiEhIWLHjh3i8OHDokePHqJHjx4yprYdf571IwTHuqHEx8cLlUol3n33XXHu3Dnx3XffCWdnZ/Htt9+az5k3b55wd3cX69atE8ePHxfDhw/nlNk6mDBhgmjWrJl5evLPP/8svL29xSuvvGI+h2Ndd/n5+SIhIUEkJCQIAGL+/PkiISFBXL58WQhRs7EdNGiQ6Nixozh48KDYu3evCA8P5/TkxvLJJ5+IkJAQ4ejoKLp16yYOHDggdySrB+CW27Jly8znFBcXi2eeeUZ4eHgIZ2dnMXLkSJGeni5faBvyv0WFY91w/vvf/4p27doJtVotIiMjxZIlS6ocN5lMYvbs2cLPz0+o1WrRr18/kZycLFNa62UwGMT06dNFSEiI0Gg0okWLFmLWrFnCaDSaz+FY191vv/12y3+jJ0yYIISo2dhev35djBs3Tri6ugqtVismTpwo8vPz651NEuJPy/oRERERWRA+o0JEREQWi0WFiIiILBaLChEREVksFhUiIiKyWCwqREREZLFYVIiIiMhisagQERGRxWJRISKrJ0kS1q5dK3cMImoELCpEVC+PP/44JEmqtg0aNEjuaERkA1RyByAi6zdo0CAsW7asyj61Wi1TGiKyJbyiQkT1plar4e/vX2Xz8PAAUHlbZuHChRg8eDCcnJzQokUL/Pjjj1W+/8SJE7j33nvh5OQELy8vTJo0CQUFBVXOWbp0KaKioqBWqxEQEIBp06ZVOZ6Tk4ORI0fC2dkZ4eHhWL9+vfnYjRs3MH78ePj4+MDJyQnh4eHVihURWSYWFSJqdLNnz8aoUaNw7NgxjB8/HmPHjkVSUhIAoLCwEAMHDoSHhwcOHTqE1atXY9u2bVWKyMKFCzF16lRMmjQJJ06cwPr169GqVasqf8ebb76J0aNH4/jx47jvvvswfvx45Obmmv/+06dPY9OmTUhKSsLChQvh7e3ddANARHVX79caEpFdmzBhglAqlcLFxaXK9u677wohKt+cPXny5CrfExsbK6ZMmSKEEGLJkiXCw8NDFBQUmI//8ssvQqFQiIyMDCGEEIGBgWLWrFm3zQBA/OMf/zB/XVBQIACITZs2CSGEGDp0qJg4cWLD/MBE1KT4jAoR1Vvfvn2xcOHCKvs8PT3N/92jR48qx3r06IHExEQAQFJSEmJiYuDi4mI+3rNnT5hMJiQnJ0OSJKSlpaFfv35/mSE6Otr83y4uLtBqtcjKygIATJkyBaNGjcLRo0cxYMAAjBgxAnfddVedflYialosKkRUby4uLtVuxTQUJyenGp3n4OBQ5WtJkmAymQAAgwcPxuXLl7Fx40Zs3boV/fr1w9SpU/Hhhx82eF4ialh8RoWIGt2BAweqfd2mTRsAQJs2bXDs2DEUFhaaj//+++9QKBSIiIiAm5sbmjdvju3bt9crg4+PDyZMmIBvv/0WCxYswJIlS+r1eUTUNHhFhYjqzWg0IiMjo8o+lUplfmB19erV6NKlC3r16oXvvvsO8fHx+OqrrwAA48ePxxtvvIEJEyZgzpw5yM7OxrPPPotHH30Ufn5+AIA5c+Zg8uTJ8PX1xeDBg5Gfn4/ff/8dzz77bI3yvf766+jcuTOioqJgNBqxYcMGc1EiIsvGokJE9bZ582YEBARU2RcREYEzZ84AqJyRs3LlSjzzzDMICAjAihUr0LZtWwCAs7MztmzZgunTp6Nr165wdnbGqFGjMH/+fPNnTZgwASUlJfjXv/6Fl156Cd7e3njwwQdrnM/R0REzZ87EpUuX4OTkhN69e2PlypUN8JMTUWOThBBC7hBEZLskScKaNWswYsQIuaMQkRXiMypERERksVhUiIiIyGLxGRUialS8u0xE9cErKkRERGSxWFSIiIjIYrGoEBERkcViUSEiIiKLxaJCREREFotFhYiIiCwWiwoRERFZLBYVIiIislgsKkRERGSx/g+qpSSRumgbDwAAAABJRU5ErkJggg==",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(model_history.history[\"loss\"])\n",
    "plt.xlabel(\"Epochs\")\n",
    "plt.ylabel(\"Loss\")\n",
    "plt.title(\"Loss vs Epoch\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "def predict_next_word(text):\n",
    "    tokenized = tokenizer.texts_to_sequences([text])[0]\n",
    "    padded = pad_sequence(tokenized, max_sequence_length)\n",
    "    input = tf.constant([padded], dtype=tf.float32)\n",
    "    prediction = model.predict(input)\n",
    "    # print(prediction[0].argmax())\n",
    "    predicted_word_index = tf.argmax(prediction, axis=1).numpy()[0]\n",
    "    \n",
    "    predicted_word = index2word[predicted_word_index]\n",
    "    print(f\"{text} : {predicted_word}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'landlocked'"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "index2word[74]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 24ms/step\n",
      "Nepal is a landlocked : country\n"
     ]
    }
   ],
   "source": [
    "predict_next_word(\"Nepal is a landlocked\")"
   ]
  }
 ],
 "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": "3.12.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}