Fine tune BERT for sentiment analysis.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>description</th>\n",
       "      <th>sentiment</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Bromwell High is a cartoon comedy. It ran at t...</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Homelessness (or Houselessness as George Carli...</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Brilliant over-acting by Lesley Ann Warren. Be...</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>This is easily the most underrated film inn th...</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>This is not the typical Mel Brooks film. It wa...</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24995</th>\n",
       "      <td>Towards the end of the movie, I felt it was to...</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24996</th>\n",
       "      <td>This is the kind of movie that my enemies cont...</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24997</th>\n",
       "      <td>I saw 'Descent' last night at the Stockholm Fi...</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24998</th>\n",
       "      <td>Some films that you pick up for a pound turn o...</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24999</th>\n",
       "      <td>This is one of the dumbest films, I've ever se...</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>25000 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                             description  sentiment\n",
       "0      Bromwell High is a cartoon comedy. It ran at t...        1.0\n",
       "1      Homelessness (or Houselessness as George Carli...        1.0\n",
       "2      Brilliant over-acting by Lesley Ann Warren. Be...        1.0\n",
       "3      This is easily the most underrated film inn th...        1.0\n",
       "4      This is not the typical Mel Brooks film. It wa...        1.0\n",
       "...                                                  ...        ...\n",
       "24995  Towards the end of the movie, I felt it was to...        0.0\n",
       "24996  This is the kind of movie that my enemies cont...        0.0\n",
       "24997  I saw 'Descent' last night at the Stockholm Fi...        0.0\n",
       "24998  Some films that you pick up for a pound turn o...        0.0\n",
       "24999  This is one of the dumbest films, I've ever se...        0.0\n",
       "\n",
       "[25000 rows x 2 columns]"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "\n",
    "df_train_sentiment = pd.read_csv(\"all_train_sentiment.csv\")\n",
    "df_train_sentiment"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>description</th>\n",
       "      <th>sentiment</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>I went and saw this movie last night after bei...</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Actor turned director Bill Paxton follows up h...</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>As a recreational golfer with some knowledge o...</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>I saw this film in a sneak preview, and it is ...</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Bill Paxton has taken the true story of the 19...</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24995</th>\n",
       "      <td>I occasionally let my kids watch this garbage ...</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24996</th>\n",
       "      <td>When all we have anymore is pretty much realit...</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24997</th>\n",
       "      <td>The basic genre is a thriller intercut with an...</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24998</th>\n",
       "      <td>Four things intrigued me as to this film - fir...</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24999</th>\n",
       "      <td>David Bryce's comments nearby are exceptionall...</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>25000 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                                             description  sentiment\n",
       "0      I went and saw this movie last night after bei...        1.0\n",
       "1      Actor turned director Bill Paxton follows up h...        1.0\n",
       "2      As a recreational golfer with some knowledge o...        1.0\n",
       "3      I saw this film in a sneak preview, and it is ...        1.0\n",
       "4      Bill Paxton has taken the true story of the 19...        1.0\n",
       "...                                                  ...        ...\n",
       "24995  I occasionally let my kids watch this garbage ...        0.0\n",
       "24996  When all we have anymore is pretty much realit...        0.0\n",
       "24997  The basic genre is a thriller intercut with an...        0.0\n",
       "24998  Four things intrigued me as to this film - fir...        0.0\n",
       "24999  David Bryce's comments nearby are exceptionall...        0.0\n",
       "\n",
       "[25000 rows x 2 columns]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_test_sentiment = pd.read_csv(\"all_test_sentiment.csv\")\n",
    "df_test_sentiment"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch\n",
    "from torch.utils.data import Dataset, DataLoader\n",
    "from transformers import BertTokenizer, BertForSequenceClassification\n",
    "from torch.optim import AdamW"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "class CustomDataset(Dataset):\n",
    "    def __init__(self, data, tokenizer, max_length):\n",
    "        self.data = data\n",
    "        self.tokenizer = tokenizer\n",
    "        self.max_length = max_length\n",
    "\n",
    "    def __len__(self):\n",
    "        return len(self.data)\n",
    "\n",
    "    def __getitem__(self, idx):\n",
    "        item = self.data.iloc[idx]\n",
    "        text = item['description']\n",
    "        label = item['sentiment']\n",
    "        inputs = self.tokenizer(text, padding='max_length', truncation=True, max_length=self.max_length, return_tensors='pt')\n",
    "        return {'input_ids': inputs['input_ids'][0], 'attention_mask': inputs['attention_mask'][0], 'labels': torch.tensor(label)}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Hyperparameters\n",
    "batch_size = 100\n",
    "max_length = 128\n",
    "learning_rate = 0.0001\n",
    "epochs = 5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Some weights of BertForSequenceClassification were not initialized from the model checkpoint at bert-base-uncased and are newly initialized: ['classifier.bias', 'classifier.weight']\n",
      "You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.\n"
     ]
    }
   ],
   "source": [
    "device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")\n",
    "\n",
    "# Initialize BERT tokenizer and model\n",
    "tokenizer = BertTokenizer.from_pretrained('bert-base-uncased')\n",
    "model = BertForSequenceClassification.from_pretrained('bert-base-uncased', num_labels=1).to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Create DataLoader for training and validation data\n",
    "train_dataset = CustomDataset(df_train_sentiment, tokenizer, max_length)\n",
    "train_loader = DataLoader(train_dataset, batch_size=batch_size, shuffle=True)\n",
    "\n",
    "test_dataset = CustomDataset(df_test_sentiment, tokenizer, max_length)\n",
    "test_loader = DataLoader(test_dataset, batch_size=batch_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define optimizer and loss function\n",
    "optimizer = AdamW(model.parameters(), lr=learning_rate)\n",
    "loss_fn = torch.nn.BCEWithLogitsLoss(reduction='none')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/5, Batch 1/250, Loss: 71.6219\n",
      "Epoch 1/5, Batch 2/250, Loss: 70.9527\n",
      "Epoch 1/5, Batch 3/250, Loss: 69.6151\n",
      "Epoch 1/5, Batch 4/250, Loss: 71.2939\n",
      "Epoch 1/5, Batch 5/250, Loss: 68.7957\n",
      "Epoch 1/5, Batch 6/250, Loss: 66.9245\n",
      "Epoch 1/5, Batch 7/250, Loss: 68.0160\n",
      "Epoch 1/5, Batch 8/250, Loss: 63.3291\n",
      "Epoch 1/5, Batch 9/250, Loss: 59.2905\n",
      "Epoch 1/5, Batch 10/250, Loss: 54.3646\n",
      "Epoch 1/5, Batch 11/250, Loss: 50.4664\n",
      "Epoch 1/5, Batch 12/250, Loss: 44.4880\n",
      "Epoch 1/5, Batch 13/250, Loss: 79.1987\n",
      "Epoch 1/5, Batch 14/250, Loss: 47.9829\n",
      "Epoch 1/5, Batch 15/250, Loss: 102.4607\n",
      "Epoch 1/5, Batch 16/250, Loss: 50.1724\n",
      "Epoch 1/5, Batch 17/250, Loss: 44.3480\n",
      "Epoch 1/5, Batch 18/250, Loss: 52.0231\n",
      "Epoch 1/5, Batch 19/250, Loss: 60.6892\n",
      "Epoch 1/5, Batch 20/250, Loss: 43.4770\n",
      "Epoch 1/5, Batch 21/250, Loss: 44.0933\n",
      "Epoch 1/5, Batch 22/250, Loss: 45.9877\n",
      "Epoch 1/5, Batch 23/250, Loss: 49.4623\n",
      "Epoch 1/5, Batch 24/250, Loss: 41.2771\n",
      "Epoch 1/5, Batch 25/250, Loss: 40.2140\n",
      "Epoch 1/5, Batch 26/250, Loss: 39.0320\n",
      "Epoch 1/5, Batch 27/250, Loss: 37.0752\n",
      "Epoch 1/5, Batch 28/250, Loss: 44.7852\n",
      "Epoch 1/5, Batch 29/250, Loss: 31.5513\n",
      "Epoch 1/5, Batch 30/250, Loss: 37.0470\n",
      "Epoch 1/5, Batch 31/250, Loss: 37.5903\n",
      "Epoch 1/5, Batch 32/250, Loss: 31.7548\n",
      "Epoch 1/5, Batch 33/250, Loss: 35.5277\n",
      "Epoch 1/5, Batch 34/250, Loss: 39.7331\n",
      "Epoch 1/5, Batch 35/250, Loss: 46.3335\n",
      "Epoch 1/5, Batch 36/250, Loss: 45.6124\n",
      "Epoch 1/5, Batch 37/250, Loss: 29.4270\n",
      "Epoch 1/5, Batch 38/250, Loss: 33.1733\n",
      "Epoch 1/5, Batch 39/250, Loss: 41.7664\n",
      "Epoch 1/5, Batch 40/250, Loss: 44.7967\n",
      "Epoch 1/5, Batch 41/250, Loss: 40.0910\n",
      "Epoch 1/5, Batch 42/250, Loss: 34.2076\n",
      "Epoch 1/5, Batch 43/250, Loss: 40.7792\n",
      "Epoch 1/5, Batch 44/250, Loss: 37.8630\n",
      "Epoch 1/5, Batch 45/250, Loss: 46.0584\n",
      "Epoch 1/5, Batch 46/250, Loss: 37.7079\n",
      "Epoch 1/5, Batch 47/250, Loss: 37.0313\n",
      "Epoch 1/5, Batch 48/250, Loss: 33.9036\n",
      "Epoch 1/5, Batch 49/250, Loss: 32.6560\n",
      "Epoch 1/5, Batch 50/250, Loss: 35.5860\n",
      "Epoch 1/5, Batch 51/250, Loss: 51.0612\n",
      "Epoch 1/5, Batch 52/250, Loss: 33.6821\n",
      "Epoch 1/5, Batch 53/250, Loss: 31.4194\n",
      "Epoch 1/5, Batch 54/250, Loss: 29.9213\n",
      "Epoch 1/5, Batch 55/250, Loss: 38.0642\n",
      "Epoch 1/5, Batch 56/250, Loss: 28.5540\n",
      "Epoch 1/5, Batch 57/250, Loss: 33.6520\n",
      "Epoch 1/5, Batch 58/250, Loss: 27.6688\n",
      "Epoch 1/5, Batch 59/250, Loss: 27.8103\n",
      "Epoch 1/5, Batch 60/250, Loss: 38.4993\n",
      "Epoch 1/5, Batch 61/250, Loss: 30.2649\n",
      "Epoch 1/5, Batch 62/250, Loss: 40.6945\n",
      "Epoch 1/5, Batch 63/250, Loss: 28.0690\n",
      "Epoch 1/5, Batch 64/250, Loss: 39.1850\n",
      "Epoch 1/5, Batch 65/250, Loss: 32.6379\n",
      "Epoch 1/5, Batch 66/250, Loss: 35.3671\n",
      "Epoch 1/5, Batch 67/250, Loss: 44.3949\n",
      "Epoch 1/5, Batch 68/250, Loss: 37.8721\n",
      "Epoch 1/5, Batch 69/250, Loss: 35.3085\n",
      "Epoch 1/5, Batch 70/250, Loss: 36.9253\n",
      "Epoch 1/5, Batch 71/250, Loss: 36.5366\n",
      "Epoch 1/5, Batch 72/250, Loss: 40.0088\n",
      "Epoch 1/5, Batch 73/250, Loss: 34.6286\n",
      "Epoch 1/5, Batch 74/250, Loss: 31.2841\n",
      "Epoch 1/5, Batch 75/250, Loss: 32.9251\n",
      "Epoch 1/5, Batch 76/250, Loss: 31.5041\n",
      "Epoch 1/5, Batch 77/250, Loss: 34.8028\n",
      "Epoch 1/5, Batch 78/250, Loss: 33.3924\n",
      "Epoch 1/5, Batch 79/250, Loss: 33.0914\n",
      "Epoch 1/5, Batch 80/250, Loss: 51.2037\n",
      "Epoch 1/5, Batch 81/250, Loss: 29.7519\n",
      "Epoch 1/5, Batch 82/250, Loss: 32.3718\n",
      "Epoch 1/5, Batch 83/250, Loss: 28.0229\n",
      "Epoch 1/5, Batch 84/250, Loss: 33.3186\n",
      "Epoch 1/5, Batch 85/250, Loss: 40.7648\n",
      "Epoch 1/5, Batch 86/250, Loss: 38.4196\n",
      "Epoch 1/5, Batch 87/250, Loss: 26.5595\n",
      "Epoch 1/5, Batch 88/250, Loss: 31.5766\n",
      "Epoch 1/5, Batch 89/250, Loss: 26.5948\n",
      "Epoch 1/5, Batch 90/250, Loss: 37.3474\n",
      "Epoch 1/5, Batch 91/250, Loss: 28.8867\n",
      "Epoch 1/5, Batch 92/250, Loss: 41.2596\n",
      "Epoch 1/5, Batch 93/250, Loss: 20.4455\n",
      "Epoch 1/5, Batch 94/250, Loss: 46.0479\n",
      "Epoch 1/5, Batch 95/250, Loss: 49.4895\n",
      "Epoch 1/5, Batch 96/250, Loss: 43.9960\n",
      "Epoch 1/5, Batch 97/250, Loss: 34.0188\n",
      "Epoch 1/5, Batch 98/250, Loss: 36.8110\n",
      "Epoch 1/5, Batch 99/250, Loss: 38.6630\n",
      "Epoch 1/5, Batch 100/250, Loss: 31.3998\n",
      "Epoch 1/5, Batch 101/250, Loss: 31.7062\n",
      "Epoch 1/5, Batch 102/250, Loss: 36.6632\n",
      "Epoch 1/5, Batch 103/250, Loss: 32.8894\n",
      "Epoch 1/5, Batch 104/250, Loss: 20.0439\n",
      "Epoch 1/5, Batch 105/250, Loss: 36.6272\n",
      "Epoch 1/5, Batch 106/250, Loss: 36.1081\n",
      "Epoch 1/5, Batch 107/250, Loss: 29.9477\n",
      "Epoch 1/5, Batch 108/250, Loss: 31.9777\n",
      "Epoch 1/5, Batch 109/250, Loss: 43.5640\n",
      "Epoch 1/5, Batch 110/250, Loss: 31.1798\n",
      "Epoch 1/5, Batch 111/250, Loss: 33.6421\n",
      "Epoch 1/5, Batch 112/250, Loss: 35.2777\n",
      "Epoch 1/5, Batch 113/250, Loss: 31.5323\n",
      "Epoch 1/5, Batch 114/250, Loss: 37.8356\n",
      "Epoch 1/5, Batch 115/250, Loss: 24.9045\n",
      "Epoch 1/5, Batch 116/250, Loss: 28.1635\n",
      "Epoch 1/5, Batch 117/250, Loss: 30.7369\n",
      "Epoch 1/5, Batch 118/250, Loss: 34.1636\n",
      "Epoch 1/5, Batch 119/250, Loss: 27.5400\n",
      "Epoch 1/5, Batch 120/250, Loss: 26.8245\n",
      "Epoch 1/5, Batch 121/250, Loss: 39.1014\n",
      "Epoch 1/5, Batch 122/250, Loss: 24.3992\n",
      "Epoch 1/5, Batch 123/250, Loss: 26.4359\n",
      "Epoch 1/5, Batch 124/250, Loss: 28.6020\n",
      "Epoch 1/5, Batch 125/250, Loss: 26.1515\n",
      "Epoch 1/5, Batch 126/250, Loss: 31.9257\n",
      "Epoch 1/5, Batch 127/250, Loss: 28.1374\n",
      "Epoch 1/5, Batch 128/250, Loss: 28.4290\n",
      "Epoch 1/5, Batch 129/250, Loss: 26.5346\n",
      "Epoch 1/5, Batch 130/250, Loss: 26.2689\n",
      "Epoch 1/5, Batch 131/250, Loss: 34.4468\n",
      "Epoch 1/5, Batch 132/250, Loss: 21.6162\n",
      "Epoch 1/5, Batch 133/250, Loss: 36.7098\n",
      "Epoch 1/5, Batch 134/250, Loss: 31.7799\n",
      "Epoch 1/5, Batch 135/250, Loss: 23.5661\n",
      "Epoch 1/5, Batch 136/250, Loss: 36.6394\n",
      "Epoch 1/5, Batch 137/250, Loss: 28.7620\n",
      "Epoch 1/5, Batch 138/250, Loss: 28.9797\n",
      "Epoch 1/5, Batch 139/250, Loss: 29.7731\n",
      "Epoch 1/5, Batch 140/250, Loss: 28.8181\n",
      "Epoch 1/5, Batch 141/250, Loss: 31.4726\n",
      "Epoch 1/5, Batch 142/250, Loss: 31.3851\n",
      "Epoch 1/5, Batch 143/250, Loss: 29.1227\n",
      "Epoch 1/5, Batch 144/250, Loss: 35.0909\n",
      "Epoch 1/5, Batch 145/250, Loss: 33.2952\n",
      "Epoch 1/5, Batch 146/250, Loss: 37.1186\n",
      "Epoch 1/5, Batch 147/250, Loss: 30.5294\n",
      "Epoch 1/5, Batch 148/250, Loss: 27.3349\n",
      "Epoch 1/5, Batch 149/250, Loss: 33.6109\n",
      "Epoch 1/5, Batch 150/250, Loss: 37.8162\n",
      "Epoch 1/5, Batch 151/250, Loss: 21.4101\n",
      "Epoch 1/5, Batch 152/250, Loss: 30.9166\n",
      "Epoch 1/5, Batch 153/250, Loss: 26.8017\n",
      "Epoch 1/5, Batch 154/250, Loss: 32.7951\n",
      "Epoch 1/5, Batch 155/250, Loss: 25.4894\n",
      "Epoch 1/5, Batch 156/250, Loss: 25.5067\n",
      "Epoch 1/5, Batch 157/250, Loss: 29.5055\n",
      "Epoch 1/5, Batch 158/250, Loss: 34.0126\n",
      "Epoch 1/5, Batch 159/250, Loss: 22.3365\n",
      "Epoch 1/5, Batch 160/250, Loss: 22.1221\n",
      "Epoch 1/5, Batch 161/250, Loss: 34.6288\n",
      "Epoch 1/5, Batch 162/250, Loss: 39.6836\n",
      "Epoch 1/5, Batch 163/250, Loss: 28.8968\n",
      "Epoch 1/5, Batch 164/250, Loss: 29.3313\n",
      "Epoch 1/5, Batch 165/250, Loss: 41.4170\n",
      "Epoch 1/5, Batch 166/250, Loss: 42.3603\n",
      "Epoch 1/5, Batch 167/250, Loss: 29.0055\n",
      "Epoch 1/5, Batch 168/250, Loss: 28.2474\n",
      "Epoch 1/5, Batch 169/250, Loss: 37.7801\n",
      "Epoch 1/5, Batch 170/250, Loss: 24.8808\n",
      "Epoch 1/5, Batch 171/250, Loss: 34.9128\n",
      "Epoch 1/5, Batch 172/250, Loss: 31.3956\n",
      "Epoch 1/5, Batch 173/250, Loss: 26.4040\n",
      "Epoch 1/5, Batch 174/250, Loss: 28.8848\n",
      "Epoch 1/5, Batch 175/250, Loss: 20.0789\n",
      "Epoch 1/5, Batch 176/250, Loss: 35.5540\n",
      "Epoch 1/5, Batch 177/250, Loss: 37.7955\n",
      "Epoch 1/5, Batch 178/250, Loss: 25.2372\n",
      "Epoch 1/5, Batch 179/250, Loss: 31.5264\n",
      "Epoch 1/5, Batch 180/250, Loss: 31.3393\n",
      "Epoch 1/5, Batch 181/250, Loss: 38.3235\n",
      "Epoch 1/5, Batch 182/250, Loss: 32.2050\n",
      "Epoch 1/5, Batch 183/250, Loss: 35.3295\n",
      "Epoch 1/5, Batch 184/250, Loss: 37.0552\n",
      "Epoch 1/5, Batch 185/250, Loss: 29.8655\n",
      "Epoch 1/5, Batch 186/250, Loss: 32.5550\n",
      "Epoch 1/5, Batch 187/250, Loss: 26.7530\n",
      "Epoch 1/5, Batch 188/250, Loss: 33.9988\n",
      "Epoch 1/5, Batch 189/250, Loss: 37.7381\n",
      "Epoch 1/5, Batch 190/250, Loss: 33.9256\n",
      "Epoch 1/5, Batch 191/250, Loss: 30.4614\n",
      "Epoch 1/5, Batch 192/250, Loss: 27.0225\n",
      "Epoch 1/5, Batch 193/250, Loss: 28.8752\n",
      "Epoch 1/5, Batch 194/250, Loss: 28.0916\n",
      "Epoch 1/5, Batch 195/250, Loss: 38.8072\n",
      "Epoch 1/5, Batch 196/250, Loss: 29.1891\n",
      "Epoch 1/5, Batch 197/250, Loss: 38.2507\n",
      "Epoch 1/5, Batch 198/250, Loss: 31.4827\n",
      "Epoch 1/5, Batch 199/250, Loss: 40.6373\n",
      "Epoch 1/5, Batch 200/250, Loss: 28.3456\n",
      "Epoch 1/5, Batch 201/250, Loss: 28.0642\n",
      "Epoch 1/5, Batch 202/250, Loss: 42.0826\n",
      "Epoch 1/5, Batch 203/250, Loss: 37.6953\n",
      "Epoch 1/5, Batch 204/250, Loss: 23.1818\n",
      "Epoch 1/5, Batch 205/250, Loss: 30.3806\n",
      "Epoch 1/5, Batch 206/250, Loss: 20.0786\n",
      "Epoch 1/5, Batch 207/250, Loss: 23.3104\n",
      "Epoch 1/5, Batch 208/250, Loss: 25.3523\n",
      "Epoch 1/5, Batch 209/250, Loss: 26.0462\n",
      "Epoch 1/5, Batch 210/250, Loss: 33.4391\n",
      "Epoch 1/5, Batch 211/250, Loss: 13.0481\n",
      "Epoch 1/5, Batch 212/250, Loss: 34.7794\n",
      "Epoch 1/5, Batch 213/250, Loss: 55.5063\n",
      "Epoch 1/5, Batch 214/250, Loss: 42.0677\n",
      "Epoch 1/5, Batch 215/250, Loss: 29.3509\n",
      "Epoch 1/5, Batch 216/250, Loss: 33.2062\n",
      "Epoch 1/5, Batch 217/250, Loss: 17.6292\n",
      "Epoch 1/5, Batch 218/250, Loss: 40.7908\n",
      "Epoch 1/5, Batch 219/250, Loss: 25.8941\n",
      "Epoch 1/5, Batch 220/250, Loss: 32.4539\n",
      "Epoch 1/5, Batch 221/250, Loss: 35.0268\n",
      "Epoch 1/5, Batch 222/250, Loss: 27.5242\n",
      "Epoch 1/5, Batch 223/250, Loss: 29.9753\n",
      "Epoch 1/5, Batch 224/250, Loss: 25.5315\n",
      "Epoch 1/5, Batch 225/250, Loss: 27.6290\n",
      "Epoch 1/5, Batch 226/250, Loss: 30.9074\n",
      "Epoch 1/5, Batch 227/250, Loss: 26.5015\n",
      "Epoch 1/5, Batch 228/250, Loss: 32.4013\n",
      "Epoch 1/5, Batch 229/250, Loss: 30.6368\n",
      "Epoch 1/5, Batch 230/250, Loss: 29.8475\n",
      "Epoch 1/5, Batch 231/250, Loss: 35.7910\n",
      "Epoch 1/5, Batch 232/250, Loss: 19.3504\n",
      "Epoch 1/5, Batch 233/250, Loss: 40.5776\n",
      "Epoch 1/5, Batch 234/250, Loss: 39.0926\n",
      "Epoch 1/5, Batch 235/250, Loss: 26.4622\n",
      "Epoch 1/5, Batch 236/250, Loss: 33.5842\n",
      "Epoch 1/5, Batch 237/250, Loss: 33.4495\n",
      "Epoch 1/5, Batch 238/250, Loss: 33.0739\n",
      "Epoch 1/5, Batch 239/250, Loss: 33.4541\n",
      "Epoch 1/5, Batch 240/250, Loss: 24.6767\n",
      "Epoch 1/5, Batch 241/250, Loss: 29.9453\n",
      "Epoch 1/5, Batch 242/250, Loss: 22.1098\n",
      "Epoch 1/5, Batch 243/250, Loss: 19.6660\n",
      "Epoch 1/5, Batch 244/250, Loss: 40.3998\n",
      "Epoch 1/5, Batch 245/250, Loss: 21.6669\n",
      "Epoch 1/5, Batch 246/250, Loss: 26.1483\n",
      "Epoch 1/5, Batch 247/250, Loss: 33.8136\n",
      "Epoch 1/5, Batch 248/250, Loss: 22.5706\n",
      "Epoch 1/5, Batch 249/250, Loss: 24.4836\n",
      "Epoch 1/5, Batch 250/250, Loss: 38.8894\n",
      "Epoch 1/5\n",
      "Training Loss: 35.2068\n",
      "Validation Loss: 27.7612\n",
      "Validation Accuracy: 0.8843\n",
      "\n",
      "Epoch 2/5, Batch 1/250, Loss: 25.2246\n",
      "Epoch 2/5, Batch 2/250, Loss: 17.3206\n",
      "Epoch 2/5, Batch 3/250, Loss: 18.5087\n",
      "Epoch 2/5, Batch 4/250, Loss: 15.5033\n",
      "Epoch 2/5, Batch 5/250, Loss: 17.9777\n",
      "Epoch 2/5, Batch 6/250, Loss: 27.6712\n",
      "Epoch 2/5, Batch 7/250, Loss: 13.7148\n",
      "Epoch 2/5, Batch 8/250, Loss: 15.3959\n",
      "Epoch 2/5, Batch 9/250, Loss: 26.4982\n",
      "Epoch 2/5, Batch 10/250, Loss: 13.3976\n",
      "Epoch 2/5, Batch 11/250, Loss: 16.8701\n",
      "Epoch 2/5, Batch 12/250, Loss: 22.0633\n",
      "Epoch 2/5, Batch 13/250, Loss: 31.0934\n",
      "Epoch 2/5, Batch 14/250, Loss: 20.9261\n",
      "Epoch 2/5, Batch 15/250, Loss: 13.0289\n",
      "Epoch 2/5, Batch 16/250, Loss: 19.3126\n",
      "Epoch 2/5, Batch 17/250, Loss: 26.7244\n",
      "Epoch 2/5, Batch 18/250, Loss: 20.1341\n",
      "Epoch 2/5, Batch 19/250, Loss: 17.9590\n",
      "Epoch 2/5, Batch 20/250, Loss: 12.4281\n",
      "Epoch 2/5, Batch 21/250, Loss: 24.9710\n",
      "Epoch 2/5, Batch 22/250, Loss: 16.9368\n",
      "Epoch 2/5, Batch 23/250, Loss: 25.0739\n",
      "Epoch 2/5, Batch 24/250, Loss: 20.2308\n",
      "Epoch 2/5, Batch 25/250, Loss: 29.7879\n",
      "Epoch 2/5, Batch 26/250, Loss: 13.9117\n",
      "Epoch 2/5, Batch 27/250, Loss: 17.1600\n",
      "Epoch 2/5, Batch 28/250, Loss: 22.3842\n",
      "Epoch 2/5, Batch 29/250, Loss: 17.9972\n",
      "Epoch 2/5, Batch 30/250, Loss: 16.2396\n",
      "Epoch 2/5, Batch 31/250, Loss: 17.2843\n",
      "Epoch 2/5, Batch 32/250, Loss: 14.5032\n",
      "Epoch 2/5, Batch 33/250, Loss: 19.8956\n",
      "Epoch 2/5, Batch 34/250, Loss: 24.9355\n",
      "Epoch 2/5, Batch 35/250, Loss: 23.4305\n",
      "Epoch 2/5, Batch 36/250, Loss: 12.4992\n",
      "Epoch 2/5, Batch 37/250, Loss: 15.6131\n",
      "Epoch 2/5, Batch 38/250, Loss: 24.0221\n",
      "Epoch 2/5, Batch 39/250, Loss: 12.2972\n",
      "Epoch 2/5, Batch 40/250, Loss: 14.5620\n",
      "Epoch 2/5, Batch 41/250, Loss: 9.9421\n",
      "Epoch 2/5, Batch 42/250, Loss: 14.9022\n",
      "Epoch 2/5, Batch 43/250, Loss: 18.8702\n",
      "Epoch 2/5, Batch 44/250, Loss: 14.6103\n",
      "Epoch 2/5, Batch 45/250, Loss: 11.5967\n",
      "Epoch 2/5, Batch 46/250, Loss: 26.3751\n",
      "Epoch 2/5, Batch 47/250, Loss: 17.9201\n",
      "Epoch 2/5, Batch 48/250, Loss: 12.4752\n",
      "Epoch 2/5, Batch 49/250, Loss: 30.4147\n",
      "Epoch 2/5, Batch 50/250, Loss: 21.0756\n",
      "Epoch 2/5, Batch 51/250, Loss: 40.2755\n",
      "Epoch 2/5, Batch 52/250, Loss: 17.0634\n",
      "Epoch 2/5, Batch 53/250, Loss: 15.2002\n",
      "Epoch 2/5, Batch 54/250, Loss: 19.0370\n",
      "Epoch 2/5, Batch 55/250, Loss: 21.4247\n",
      "Epoch 2/5, Batch 56/250, Loss: 19.4035\n",
      "Epoch 2/5, Batch 57/250, Loss: 18.0034\n",
      "Epoch 2/5, Batch 58/250, Loss: 20.7670\n",
      "Epoch 2/5, Batch 59/250, Loss: 16.0366\n",
      "Epoch 2/5, Batch 60/250, Loss: 14.2856\n",
      "Epoch 2/5, Batch 61/250, Loss: 15.5989\n",
      "Epoch 2/5, Batch 62/250, Loss: 21.3635\n",
      "Epoch 2/5, Batch 63/250, Loss: 22.7693\n",
      "Epoch 2/5, Batch 64/250, Loss: 14.4751\n",
      "Epoch 2/5, Batch 65/250, Loss: 17.4045\n",
      "Epoch 2/5, Batch 66/250, Loss: 28.6773\n",
      "Epoch 2/5, Batch 67/250, Loss: 15.6947\n",
      "Epoch 2/5, Batch 68/250, Loss: 18.3198\n",
      "Epoch 2/5, Batch 69/250, Loss: 13.6947\n",
      "Epoch 2/5, Batch 70/250, Loss: 24.6812\n",
      "Epoch 2/5, Batch 71/250, Loss: 12.8887\n",
      "Epoch 2/5, Batch 72/250, Loss: 26.1386\n",
      "Epoch 2/5, Batch 73/250, Loss: 27.3758\n",
      "Epoch 2/5, Batch 74/250, Loss: 19.1505\n",
      "Epoch 2/5, Batch 75/250, Loss: 19.6638\n",
      "Epoch 2/5, Batch 76/250, Loss: 13.0377\n",
      "Epoch 2/5, Batch 77/250, Loss: 21.2881\n",
      "Epoch 2/5, Batch 78/250, Loss: 17.6018\n",
      "Epoch 2/5, Batch 79/250, Loss: 17.9941\n",
      "Epoch 2/5, Batch 80/250, Loss: 16.9180\n",
      "Epoch 2/5, Batch 81/250, Loss: 16.4632\n",
      "Epoch 2/5, Batch 82/250, Loss: 16.6801\n",
      "Epoch 2/5, Batch 83/250, Loss: 26.8145\n",
      "Epoch 2/5, Batch 84/250, Loss: 20.4494\n",
      "Epoch 2/5, Batch 85/250, Loss: 13.4691\n",
      "Epoch 2/5, Batch 86/250, Loss: 17.9405\n",
      "Epoch 2/5, Batch 87/250, Loss: 18.5649\n",
      "Epoch 2/5, Batch 88/250, Loss: 9.5485\n",
      "Epoch 2/5, Batch 89/250, Loss: 13.9748\n",
      "Epoch 2/5, Batch 90/250, Loss: 16.0615\n",
      "Epoch 2/5, Batch 91/250, Loss: 19.4426\n",
      "Epoch 2/5, Batch 92/250, Loss: 15.1726\n",
      "Epoch 2/5, Batch 93/250, Loss: 21.2779\n",
      "Epoch 2/5, Batch 94/250, Loss: 44.8381\n",
      "Epoch 2/5, Batch 95/250, Loss: 15.0847\n",
      "Epoch 2/5, Batch 96/250, Loss: 37.0890\n",
      "Epoch 2/5, Batch 97/250, Loss: 30.1902\n",
      "Epoch 2/5, Batch 98/250, Loss: 18.3144\n",
      "Epoch 2/5, Batch 99/250, Loss: 29.1174\n",
      "Epoch 2/5, Batch 100/250, Loss: 24.9051\n",
      "Epoch 2/5, Batch 101/250, Loss: 21.4144\n",
      "Epoch 2/5, Batch 102/250, Loss: 14.0335\n",
      "Epoch 2/5, Batch 103/250, Loss: 17.5988\n",
      "Epoch 2/5, Batch 104/250, Loss: 23.7254\n",
      "Epoch 2/5, Batch 105/250, Loss: 24.8544\n",
      "Epoch 2/5, Batch 106/250, Loss: 32.7088\n",
      "Epoch 2/5, Batch 107/250, Loss: 14.9858\n",
      "Epoch 2/5, Batch 108/250, Loss: 17.3337\n",
      "Epoch 2/5, Batch 109/250, Loss: 19.9604\n",
      "Epoch 2/5, Batch 110/250, Loss: 23.2172\n",
      "Epoch 2/5, Batch 111/250, Loss: 14.2940\n",
      "Epoch 2/5, Batch 112/250, Loss: 10.0545\n",
      "Epoch 2/5, Batch 113/250, Loss: 17.0762\n",
      "Epoch 2/5, Batch 114/250, Loss: 21.7805\n",
      "Epoch 2/5, Batch 115/250, Loss: 26.4036\n",
      "Epoch 2/5, Batch 116/250, Loss: 21.1693\n",
      "Epoch 2/5, Batch 117/250, Loss: 19.9305\n",
      "Epoch 2/5, Batch 118/250, Loss: 13.3593\n",
      "Epoch 2/5, Batch 119/250, Loss: 14.9133\n",
      "Epoch 2/5, Batch 120/250, Loss: 21.7494\n",
      "Epoch 2/5, Batch 121/250, Loss: 19.3928\n",
      "Epoch 2/5, Batch 122/250, Loss: 15.5125\n",
      "Epoch 2/5, Batch 123/250, Loss: 14.7688\n",
      "Epoch 2/5, Batch 124/250, Loss: 23.1717\n",
      "Epoch 2/5, Batch 125/250, Loss: 31.6767\n",
      "Epoch 2/5, Batch 126/250, Loss: 20.4256\n",
      "Epoch 2/5, Batch 127/250, Loss: 15.8166\n",
      "Epoch 2/5, Batch 128/250, Loss: 16.1681\n",
      "Epoch 2/5, Batch 129/250, Loss: 17.5278\n",
      "Epoch 2/5, Batch 130/250, Loss: 17.2791\n",
      "Epoch 2/5, Batch 131/250, Loss: 19.3476\n",
      "Epoch 2/5, Batch 132/250, Loss: 17.9254\n",
      "Epoch 2/5, Batch 133/250, Loss: 23.4626\n",
      "Epoch 2/5, Batch 134/250, Loss: 20.6597\n",
      "Epoch 2/5, Batch 135/250, Loss: 22.1465\n",
      "Epoch 2/5, Batch 136/250, Loss: 29.3912\n",
      "Epoch 2/5, Batch 137/250, Loss: 21.8677\n",
      "Epoch 2/5, Batch 138/250, Loss: 19.4174\n",
      "Epoch 2/5, Batch 139/250, Loss: 24.8755\n",
      "Epoch 2/5, Batch 140/250, Loss: 25.5601\n",
      "Epoch 2/5, Batch 141/250, Loss: 36.3387\n",
      "Epoch 2/5, Batch 142/250, Loss: 15.7946\n",
      "Epoch 2/5, Batch 143/250, Loss: 34.6347\n",
      "Epoch 2/5, Batch 144/250, Loss: 27.0484\n",
      "Epoch 2/5, Batch 145/250, Loss: 20.1839\n",
      "Epoch 2/5, Batch 146/250, Loss: 18.4407\n",
      "Epoch 2/5, Batch 147/250, Loss: 20.4784\n",
      "Epoch 2/5, Batch 148/250, Loss: 15.5178\n",
      "Epoch 2/5, Batch 149/250, Loss: 26.5432\n",
      "Epoch 2/5, Batch 150/250, Loss: 16.2755\n",
      "Epoch 2/5, Batch 151/250, Loss: 14.7693\n",
      "Epoch 2/5, Batch 152/250, Loss: 22.4698\n",
      "Epoch 2/5, Batch 153/250, Loss: 15.7336\n",
      "Epoch 2/5, Batch 154/250, Loss: 14.3594\n",
      "Epoch 2/5, Batch 155/250, Loss: 11.3043\n",
      "Epoch 2/5, Batch 156/250, Loss: 16.8336\n",
      "Epoch 2/5, Batch 157/250, Loss: 22.6886\n",
      "Epoch 2/5, Batch 158/250, Loss: 19.2622\n",
      "Epoch 2/5, Batch 159/250, Loss: 22.3177\n",
      "Epoch 2/5, Batch 160/250, Loss: 10.8234\n",
      "Epoch 2/5, Batch 161/250, Loss: 20.7941\n",
      "Epoch 2/5, Batch 162/250, Loss: 19.3736\n",
      "Epoch 2/5, Batch 163/250, Loss: 25.2525\n",
      "Epoch 2/5, Batch 164/250, Loss: 15.8768\n",
      "Epoch 2/5, Batch 165/250, Loss: 20.6925\n",
      "Epoch 2/5, Batch 166/250, Loss: 11.4424\n",
      "Epoch 2/5, Batch 167/250, Loss: 20.6122\n",
      "Epoch 2/5, Batch 168/250, Loss: 17.8955\n",
      "Epoch 2/5, Batch 169/250, Loss: 19.5037\n",
      "Epoch 2/5, Batch 170/250, Loss: 12.1909\n",
      "Epoch 2/5, Batch 171/250, Loss: 14.6604\n",
      "Epoch 2/5, Batch 172/250, Loss: 17.8912\n",
      "Epoch 2/5, Batch 173/250, Loss: 28.6693\n",
      "Epoch 2/5, Batch 174/250, Loss: 14.1834\n",
      "Epoch 2/5, Batch 175/250, Loss: 18.4940\n",
      "Epoch 2/5, Batch 176/250, Loss: 13.6129\n",
      "Epoch 2/5, Batch 177/250, Loss: 25.2281\n",
      "Epoch 2/5, Batch 178/250, Loss: 20.4919\n",
      "Epoch 2/5, Batch 179/250, Loss: 22.9900\n",
      "Epoch 2/5, Batch 180/250, Loss: 13.2454\n",
      "Epoch 2/5, Batch 181/250, Loss: 19.8037\n",
      "Epoch 2/5, Batch 182/250, Loss: 19.3764\n",
      "Epoch 2/5, Batch 183/250, Loss: 11.6544\n",
      "Epoch 2/5, Batch 184/250, Loss: 9.3173\n",
      "Epoch 2/5, Batch 185/250, Loss: 15.4625\n",
      "Epoch 2/5, Batch 186/250, Loss: 17.9702\n",
      "Epoch 2/5, Batch 187/250, Loss: 15.9928\n",
      "Epoch 2/5, Batch 188/250, Loss: 20.6999\n",
      "Epoch 2/5, Batch 189/250, Loss: 13.5659\n",
      "Epoch 2/5, Batch 190/250, Loss: 20.7103\n",
      "Epoch 2/5, Batch 191/250, Loss: 14.9239\n",
      "Epoch 2/5, Batch 192/250, Loss: 14.3168\n",
      "Epoch 2/5, Batch 193/250, Loss: 20.1806\n",
      "Epoch 2/5, Batch 194/250, Loss: 23.3778\n",
      "Epoch 2/5, Batch 195/250, Loss: 12.6850\n",
      "Epoch 2/5, Batch 196/250, Loss: 26.4942\n",
      "Epoch 2/5, Batch 197/250, Loss: 14.3507\n",
      "Epoch 2/5, Batch 198/250, Loss: 15.3051\n",
      "Epoch 2/5, Batch 199/250, Loss: 20.3763\n",
      "Epoch 2/5, Batch 200/250, Loss: 25.5412\n",
      "Epoch 2/5, Batch 201/250, Loss: 10.9970\n",
      "Epoch 2/5, Batch 202/250, Loss: 24.1821\n",
      "Epoch 2/5, Batch 203/250, Loss: 15.8636\n",
      "Epoch 2/5, Batch 204/250, Loss: 20.3847\n",
      "Epoch 2/5, Batch 205/250, Loss: 16.6143\n",
      "Epoch 2/5, Batch 206/250, Loss: 14.4977\n",
      "Epoch 2/5, Batch 207/250, Loss: 15.9879\n",
      "Epoch 2/5, Batch 208/250, Loss: 15.1005\n",
      "Epoch 2/5, Batch 209/250, Loss: 24.5286\n",
      "Epoch 2/5, Batch 210/250, Loss: 19.4093\n",
      "Epoch 2/5, Batch 211/250, Loss: 17.6542\n",
      "Epoch 2/5, Batch 212/250, Loss: 30.8088\n",
      "Epoch 2/5, Batch 213/250, Loss: 10.5116\n",
      "Epoch 2/5, Batch 214/250, Loss: 20.4471\n",
      "Epoch 2/5, Batch 215/250, Loss: 13.4358\n",
      "Epoch 2/5, Batch 216/250, Loss: 22.3033\n",
      "Epoch 2/5, Batch 217/250, Loss: 16.3945\n",
      "Epoch 2/5, Batch 218/250, Loss: 12.0390\n",
      "Epoch 2/5, Batch 219/250, Loss: 13.8991\n",
      "Epoch 2/5, Batch 220/250, Loss: 13.6830\n",
      "Epoch 2/5, Batch 221/250, Loss: 16.3528\n",
      "Epoch 2/5, Batch 222/250, Loss: 16.1081\n",
      "Epoch 2/5, Batch 223/250, Loss: 22.2562\n",
      "Epoch 2/5, Batch 224/250, Loss: 14.9548\n",
      "Epoch 2/5, Batch 225/250, Loss: 30.5657\n",
      "Epoch 2/5, Batch 226/250, Loss: 20.5032\n",
      "Epoch 2/5, Batch 227/250, Loss: 18.5803\n",
      "Epoch 2/5, Batch 228/250, Loss: 14.2725\n",
      "Epoch 2/5, Batch 229/250, Loss: 18.6955\n",
      "Epoch 2/5, Batch 230/250, Loss: 20.5720\n",
      "Epoch 2/5, Batch 231/250, Loss: 13.2861\n",
      "Epoch 2/5, Batch 232/250, Loss: 20.6358\n",
      "Epoch 2/5, Batch 233/250, Loss: 10.7486\n",
      "Epoch 2/5, Batch 234/250, Loss: 17.2346\n",
      "Epoch 2/5, Batch 235/250, Loss: 11.4158\n",
      "Epoch 2/5, Batch 236/250, Loss: 12.8968\n",
      "Epoch 2/5, Batch 237/250, Loss: 19.4807\n",
      "Epoch 2/5, Batch 238/250, Loss: 23.5877\n",
      "Epoch 2/5, Batch 239/250, Loss: 20.1892\n",
      "Epoch 2/5, Batch 240/250, Loss: 22.4076\n",
      "Epoch 2/5, Batch 241/250, Loss: 25.1890\n",
      "Epoch 2/5, Batch 242/250, Loss: 16.1175\n",
      "Epoch 2/5, Batch 243/250, Loss: 25.1227\n",
      "Epoch 2/5, Batch 244/250, Loss: 22.3892\n",
      "Epoch 2/5, Batch 245/250, Loss: 14.5694\n",
      "Epoch 2/5, Batch 246/250, Loss: 19.4104\n",
      "Epoch 2/5, Batch 247/250, Loss: 17.2779\n",
      "Epoch 2/5, Batch 248/250, Loss: 16.8648\n",
      "Epoch 2/5, Batch 249/250, Loss: 14.7878\n",
      "Epoch 2/5, Batch 250/250, Loss: 17.1042\n",
      "Epoch 2/5\n",
      "Training Loss: 19.2064\n",
      "Validation Loss: 32.0721\n",
      "Validation Accuracy: 0.8734\n",
      "\n",
      "Epoch 3/5, Batch 1/250, Loss: 8.9605\n",
      "Epoch 3/5, Batch 2/250, Loss: 8.9042\n",
      "Epoch 3/5, Batch 3/250, Loss: 8.0699\n",
      "Epoch 3/5, Batch 4/250, Loss: 5.9028\n",
      "Epoch 3/5, Batch 5/250, Loss: 8.4224\n",
      "Epoch 3/5, Batch 6/250, Loss: 3.3715\n",
      "Epoch 3/5, Batch 7/250, Loss: 17.0552\n",
      "Epoch 3/5, Batch 8/250, Loss: 4.9561\n",
      "Epoch 3/5, Batch 9/250, Loss: 9.9887\n",
      "Epoch 3/5, Batch 10/250, Loss: 1.6496\n",
      "Epoch 3/5, Batch 11/250, Loss: 9.5903\n",
      "Epoch 3/5, Batch 12/250, Loss: 5.6589\n",
      "Epoch 3/5, Batch 13/250, Loss: 4.7305\n",
      "Epoch 3/5, Batch 14/250, Loss: 1.4969\n",
      "Epoch 3/5, Batch 15/250, Loss: 5.8416\n",
      "Epoch 3/5, Batch 16/250, Loss: 10.6308\n",
      "Epoch 3/5, Batch 17/250, Loss: 0.4066\n",
      "Epoch 3/5, Batch 18/250, Loss: 5.4468\n",
      "Epoch 3/5, Batch 19/250, Loss: 6.8476\n",
      "Epoch 3/5, Batch 20/250, Loss: 1.8720\n",
      "Epoch 3/5, Batch 21/250, Loss: 6.8661\n",
      "Epoch 3/5, Batch 22/250, Loss: 8.8088\n",
      "Epoch 3/5, Batch 23/250, Loss: 6.8897\n",
      "Epoch 3/5, Batch 24/250, Loss: 7.7215\n",
      "Epoch 3/5, Batch 25/250, Loss: 8.9528\n",
      "Epoch 3/5, Batch 26/250, Loss: 9.5266\n",
      "Epoch 3/5, Batch 27/250, Loss: 5.0309\n",
      "Epoch 3/5, Batch 28/250, Loss: 14.8339\n",
      "Epoch 3/5, Batch 29/250, Loss: 12.6280\n",
      "Epoch 3/5, Batch 30/250, Loss: 9.1058\n",
      "Epoch 3/5, Batch 31/250, Loss: 7.4532\n",
      "Epoch 3/5, Batch 32/250, Loss: 19.2142\n",
      "Epoch 3/5, Batch 33/250, Loss: 6.7879\n",
      "Epoch 3/5, Batch 34/250, Loss: 5.0297\n",
      "Epoch 3/5, Batch 35/250, Loss: 3.7047\n",
      "Epoch 3/5, Batch 36/250, Loss: 12.1598\n",
      "Epoch 3/5, Batch 37/250, Loss: 7.7008\n",
      "Epoch 3/5, Batch 38/250, Loss: 8.7206\n",
      "Epoch 3/5, Batch 39/250, Loss: 5.0160\n",
      "Epoch 3/5, Batch 40/250, Loss: 10.5343\n",
      "Epoch 3/5, Batch 41/250, Loss: 7.2932\n",
      "Epoch 3/5, Batch 42/250, Loss: 4.9817\n",
      "Epoch 3/5, Batch 43/250, Loss: 8.3111\n",
      "Epoch 3/5, Batch 44/250, Loss: 4.8473\n",
      "Epoch 3/5, Batch 45/250, Loss: 4.1945\n",
      "Epoch 3/5, Batch 46/250, Loss: 6.5342\n",
      "Epoch 3/5, Batch 47/250, Loss: 13.5059\n",
      "Epoch 3/5, Batch 48/250, Loss: 8.1991\n",
      "Epoch 3/5, Batch 49/250, Loss: 3.4463\n",
      "Epoch 3/5, Batch 50/250, Loss: 16.0506\n",
      "Epoch 3/5, Batch 51/250, Loss: 15.8288\n",
      "Epoch 3/5, Batch 52/250, Loss: 13.6056\n",
      "Epoch 3/5, Batch 53/250, Loss: 7.6400\n",
      "Epoch 3/5, Batch 54/250, Loss: 4.8837\n",
      "Epoch 3/5, Batch 55/250, Loss: 13.6436\n",
      "Epoch 3/5, Batch 56/250, Loss: 6.4553\n",
      "Epoch 3/5, Batch 57/250, Loss: 5.4234\n",
      "Epoch 3/5, Batch 58/250, Loss: 5.4618\n",
      "Epoch 3/5, Batch 59/250, Loss: 12.4551\n",
      "Epoch 3/5, Batch 60/250, Loss: 4.4508\n",
      "Epoch 3/5, Batch 61/250, Loss: 6.2907\n",
      "Epoch 3/5, Batch 62/250, Loss: 3.7975\n",
      "Epoch 3/5, Batch 63/250, Loss: 7.7077\n",
      "Epoch 3/5, Batch 64/250, Loss: 3.6376\n",
      "Epoch 3/5, Batch 65/250, Loss: 5.2912\n",
      "Epoch 3/5, Batch 66/250, Loss: 3.7303\n",
      "Epoch 3/5, Batch 67/250, Loss: 5.9275\n",
      "Epoch 3/5, Batch 68/250, Loss: 5.7185\n",
      "Epoch 3/5, Batch 69/250, Loss: 14.1854\n",
      "Epoch 3/5, Batch 70/250, Loss: 6.8494\n",
      "Epoch 3/5, Batch 71/250, Loss: 4.7980\n",
      "Epoch 3/5, Batch 72/250, Loss: 9.5629\n",
      "Epoch 3/5, Batch 73/250, Loss: 1.3445\n",
      "Epoch 3/5, Batch 74/250, Loss: 13.3908\n",
      "Epoch 3/5, Batch 75/250, Loss: 5.8863\n",
      "Epoch 3/5, Batch 76/250, Loss: 12.0588\n",
      "Epoch 3/5, Batch 77/250, Loss: 9.4574\n",
      "Epoch 3/5, Batch 78/250, Loss: 3.4557\n",
      "Epoch 3/5, Batch 79/250, Loss: 6.6221\n",
      "Epoch 3/5, Batch 80/250, Loss: 6.2333\n",
      "Epoch 3/5, Batch 81/250, Loss: 6.9565\n",
      "Epoch 3/5, Batch 82/250, Loss: 4.4720\n",
      "Epoch 3/5, Batch 83/250, Loss: 10.9226\n",
      "Epoch 3/5, Batch 84/250, Loss: 21.9035\n",
      "Epoch 3/5, Batch 85/250, Loss: 6.3256\n",
      "Epoch 3/5, Batch 86/250, Loss: 5.4053\n",
      "Epoch 3/5, Batch 87/250, Loss: 14.4411\n",
      "Epoch 3/5, Batch 88/250, Loss: 5.8928\n",
      "Epoch 3/5, Batch 89/250, Loss: 7.2719\n",
      "Epoch 3/5, Batch 90/250, Loss: 4.9392\n",
      "Epoch 3/5, Batch 91/250, Loss: 21.5811\n",
      "Epoch 3/5, Batch 92/250, Loss: 8.5974\n",
      "Epoch 3/5, Batch 93/250, Loss: 7.4296\n",
      "Epoch 3/5, Batch 94/250, Loss: 9.4896\n",
      "Epoch 3/5, Batch 95/250, Loss: 5.7640\n",
      "Epoch 3/5, Batch 96/250, Loss: 4.9676\n",
      "Epoch 3/5, Batch 97/250, Loss: 6.5733\n",
      "Epoch 3/5, Batch 98/250, Loss: 9.8482\n",
      "Epoch 3/5, Batch 99/250, Loss: 7.9929\n",
      "Epoch 3/5, Batch 100/250, Loss: 5.0140\n",
      "Epoch 3/5, Batch 101/250, Loss: 5.5758\n",
      "Epoch 3/5, Batch 102/250, Loss: 7.0985\n",
      "Epoch 3/5, Batch 103/250, Loss: 11.9038\n",
      "Epoch 3/5, Batch 104/250, Loss: 9.9143\n",
      "Epoch 3/5, Batch 105/250, Loss: 10.6421\n",
      "Epoch 3/5, Batch 106/250, Loss: 4.9856\n",
      "Epoch 3/5, Batch 107/250, Loss: 9.0007\n",
      "Epoch 3/5, Batch 108/250, Loss: 20.0931\n",
      "Epoch 3/5, Batch 109/250, Loss: 3.3818\n",
      "Epoch 3/5, Batch 110/250, Loss: 8.0905\n",
      "Epoch 3/5, Batch 111/250, Loss: 10.3354\n",
      "Epoch 3/5, Batch 112/250, Loss: 16.5616\n",
      "Epoch 3/5, Batch 113/250, Loss: 5.8038\n",
      "Epoch 3/5, Batch 114/250, Loss: 9.2847\n",
      "Epoch 3/5, Batch 115/250, Loss: 14.6717\n",
      "Epoch 3/5, Batch 116/250, Loss: 9.0442\n",
      "Epoch 3/5, Batch 117/250, Loss: 8.6794\n",
      "Epoch 3/5, Batch 118/250, Loss: 6.1811\n",
      "Epoch 3/5, Batch 119/250, Loss: 20.5011\n",
      "Epoch 3/5, Batch 120/250, Loss: 8.8657\n",
      "Epoch 3/5, Batch 121/250, Loss: 4.7191\n",
      "Epoch 3/5, Batch 122/250, Loss: 3.7020\n",
      "Epoch 3/5, Batch 123/250, Loss: 4.1165\n",
      "Epoch 3/5, Batch 124/250, Loss: 4.8169\n",
      "Epoch 3/5, Batch 125/250, Loss: 4.7230\n",
      "Epoch 3/5, Batch 126/250, Loss: 7.1716\n",
      "Epoch 3/5, Batch 127/250, Loss: 13.4639\n",
      "Epoch 3/5, Batch 128/250, Loss: 14.1863\n",
      "Epoch 3/5, Batch 129/250, Loss: 27.5174\n",
      "Epoch 3/5, Batch 130/250, Loss: 14.2307\n",
      "Epoch 3/5, Batch 131/250, Loss: 10.2213\n",
      "Epoch 3/5, Batch 132/250, Loss: 3.8049\n",
      "Epoch 3/5, Batch 133/250, Loss: 7.9757\n",
      "Epoch 3/5, Batch 134/250, Loss: 38.4771\n",
      "Epoch 3/5, Batch 135/250, Loss: 13.0258\n",
      "Epoch 3/5, Batch 136/250, Loss: 7.5658\n",
      "Epoch 3/5, Batch 137/250, Loss: 23.6445\n",
      "Epoch 3/5, Batch 138/250, Loss: 9.8094\n",
      "Epoch 3/5, Batch 139/250, Loss: 20.8539\n",
      "Epoch 3/5, Batch 140/250, Loss: 11.5533\n",
      "Epoch 3/5, Batch 141/250, Loss: 15.4599\n",
      "Epoch 3/5, Batch 142/250, Loss: 12.1031\n",
      "Epoch 3/5, Batch 143/250, Loss: 11.5309\n",
      "Epoch 3/5, Batch 144/250, Loss: 9.3543\n",
      "Epoch 3/5, Batch 145/250, Loss: 6.9946\n",
      "Epoch 3/5, Batch 146/250, Loss: 4.9084\n",
      "Epoch 3/5, Batch 147/250, Loss: 7.3224\n",
      "Epoch 3/5, Batch 148/250, Loss: 11.3631\n",
      "Epoch 3/5, Batch 149/250, Loss: 11.1617\n",
      "Epoch 3/5, Batch 150/250, Loss: 9.4280\n",
      "Epoch 3/5, Batch 151/250, Loss: 2.4966\n",
      "Epoch 3/5, Batch 152/250, Loss: 11.4098\n",
      "Epoch 3/5, Batch 153/250, Loss: 4.0058\n",
      "Epoch 3/5, Batch 154/250, Loss: 9.8578\n",
      "Epoch 3/5, Batch 155/250, Loss: 6.3949\n",
      "Epoch 3/5, Batch 156/250, Loss: 3.9086\n",
      "Epoch 3/5, Batch 157/250, Loss: 2.6845\n",
      "Epoch 3/5, Batch 158/250, Loss: 9.5476\n",
      "Epoch 3/5, Batch 159/250, Loss: 11.4311\n",
      "Epoch 3/5, Batch 160/250, Loss: 3.4986\n",
      "Epoch 3/5, Batch 161/250, Loss: 8.9568\n",
      "Epoch 3/5, Batch 162/250, Loss: 3.8497\n",
      "Epoch 3/5, Batch 163/250, Loss: 17.5415\n",
      "Epoch 3/5, Batch 164/250, Loss: 3.0082\n",
      "Epoch 3/5, Batch 165/250, Loss: 12.8843\n",
      "Epoch 3/5, Batch 166/250, Loss: 22.9762\n",
      "Epoch 3/5, Batch 167/250, Loss: 11.6022\n",
      "Epoch 3/5, Batch 168/250, Loss: 17.5394\n",
      "Epoch 3/5, Batch 169/250, Loss: 13.2728\n",
      "Epoch 3/5, Batch 170/250, Loss: 8.0078\n",
      "Epoch 3/5, Batch 171/250, Loss: 9.2985\n",
      "Epoch 3/5, Batch 172/250, Loss: 19.2940\n",
      "Epoch 3/5, Batch 173/250, Loss: 8.6225\n",
      "Epoch 3/5, Batch 174/250, Loss: 10.5339\n",
      "Epoch 3/5, Batch 175/250, Loss: 17.3246\n",
      "Epoch 3/5, Batch 176/250, Loss: 9.7511\n",
      "Epoch 3/5, Batch 177/250, Loss: 14.7047\n",
      "Epoch 3/5, Batch 178/250, Loss: 15.9882\n",
      "Epoch 3/5, Batch 179/250, Loss: 8.6020\n",
      "Epoch 3/5, Batch 180/250, Loss: 3.9570\n",
      "Epoch 3/5, Batch 181/250, Loss: 10.1126\n",
      "Epoch 3/5, Batch 182/250, Loss: 8.6521\n",
      "Epoch 3/5, Batch 183/250, Loss: 17.3825\n",
      "Epoch 3/5, Batch 184/250, Loss: 2.4920\n",
      "Epoch 3/5, Batch 185/250, Loss: 13.1069\n",
      "Epoch 3/5, Batch 186/250, Loss: 9.9496\n",
      "Epoch 3/5, Batch 187/250, Loss: 8.2162\n",
      "Epoch 3/5, Batch 188/250, Loss: 7.7203\n",
      "Epoch 3/5, Batch 189/250, Loss: 8.7234\n",
      "Epoch 3/5, Batch 190/250, Loss: 8.8553\n",
      "Epoch 3/5, Batch 191/250, Loss: 6.0949\n",
      "Epoch 3/5, Batch 192/250, Loss: 10.2174\n",
      "Epoch 3/5, Batch 193/250, Loss: 4.9474\n",
      "Epoch 3/5, Batch 194/250, Loss: 23.5596\n",
      "Epoch 3/5, Batch 195/250, Loss: 8.6809\n",
      "Epoch 3/5, Batch 196/250, Loss: 25.5772\n",
      "Epoch 3/5, Batch 197/250, Loss: 18.5200\n",
      "Epoch 3/5, Batch 198/250, Loss: 6.4927\n",
      "Epoch 3/5, Batch 199/250, Loss: 8.5863\n",
      "Epoch 3/5, Batch 200/250, Loss: 13.0794\n",
      "Epoch 3/5, Batch 201/250, Loss: 12.6447\n",
      "Epoch 3/5, Batch 202/250, Loss: 15.0964\n",
      "Epoch 3/5, Batch 203/250, Loss: 12.5780\n",
      "Epoch 3/5, Batch 204/250, Loss: 15.9895\n",
      "Epoch 3/5, Batch 205/250, Loss: 10.6178\n",
      "Epoch 3/5, Batch 206/250, Loss: 17.7754\n",
      "Epoch 3/5, Batch 207/250, Loss: 8.7522\n",
      "Epoch 3/5, Batch 208/250, Loss: 8.4050\n",
      "Epoch 3/5, Batch 209/250, Loss: 7.6994\n",
      "Epoch 3/5, Batch 210/250, Loss: 9.6878\n",
      "Epoch 3/5, Batch 211/250, Loss: 9.6252\n",
      "Epoch 3/5, Batch 212/250, Loss: 15.7022\n",
      "Epoch 3/5, Batch 213/250, Loss: 16.5402\n",
      "Epoch 3/5, Batch 214/250, Loss: 12.9148\n",
      "Epoch 3/5, Batch 215/250, Loss: 13.9140\n",
      "Epoch 3/5, Batch 216/250, Loss: 12.1872\n",
      "Epoch 3/5, Batch 217/250, Loss: 6.9134\n",
      "Epoch 3/5, Batch 218/250, Loss: 18.6284\n",
      "Epoch 3/5, Batch 219/250, Loss: 13.9690\n",
      "Epoch 3/5, Batch 220/250, Loss: 9.5955\n",
      "Epoch 3/5, Batch 221/250, Loss: 12.1131\n",
      "Epoch 3/5, Batch 222/250, Loss: 7.4642\n",
      "Epoch 3/5, Batch 223/250, Loss: 7.1678\n",
      "Epoch 3/5, Batch 224/250, Loss: 16.6810\n",
      "Epoch 3/5, Batch 225/250, Loss: 5.6381\n",
      "Epoch 3/5, Batch 226/250, Loss: 3.8763\n",
      "Epoch 3/5, Batch 227/250, Loss: 4.2101\n",
      "Epoch 3/5, Batch 228/250, Loss: 4.4143\n",
      "Epoch 3/5, Batch 229/250, Loss: 16.3780\n",
      "Epoch 3/5, Batch 230/250, Loss: 7.2788\n",
      "Epoch 3/5, Batch 231/250, Loss: 14.4384\n",
      "Epoch 3/5, Batch 232/250, Loss: 12.2719\n",
      "Epoch 3/5, Batch 233/250, Loss: 8.0374\n",
      "Epoch 3/5, Batch 234/250, Loss: 19.6856\n",
      "Epoch 3/5, Batch 235/250, Loss: 5.0242\n",
      "Epoch 3/5, Batch 236/250, Loss: 7.5372\n",
      "Epoch 3/5, Batch 237/250, Loss: 12.1763\n",
      "Epoch 3/5, Batch 238/250, Loss: 10.0131\n",
      "Epoch 3/5, Batch 239/250, Loss: 13.9392\n",
      "Epoch 3/5, Batch 240/250, Loss: 10.9193\n",
      "Epoch 3/5, Batch 241/250, Loss: 3.2782\n",
      "Epoch 3/5, Batch 242/250, Loss: 14.5436\n",
      "Epoch 3/5, Batch 243/250, Loss: 9.5111\n",
      "Epoch 3/5, Batch 244/250, Loss: 3.3992\n",
      "Epoch 3/5, Batch 245/250, Loss: 6.8464\n",
      "Epoch 3/5, Batch 246/250, Loss: 5.6682\n",
      "Epoch 3/5, Batch 247/250, Loss: 22.4915\n",
      "Epoch 3/5, Batch 248/250, Loss: 5.3157\n",
      "Epoch 3/5, Batch 249/250, Loss: 5.7185\n",
      "Epoch 3/5, Batch 250/250, Loss: 5.0720\n",
      "Epoch 3/5\n",
      "Training Loss: 9.5421\n",
      "Validation Loss: 39.2078\n",
      "Validation Accuracy: 0.8781\n",
      "\n",
      "Epoch 4/5, Batch 1/250, Loss: 1.5028\n",
      "Epoch 4/5, Batch 2/250, Loss: 8.3187\n",
      "Epoch 4/5, Batch 3/250, Loss: 1.6744\n",
      "Epoch 4/5, Batch 4/250, Loss: 3.0845\n",
      "Epoch 4/5, Batch 5/250, Loss: 2.2003\n",
      "Epoch 4/5, Batch 6/250, Loss: 6.4025\n",
      "Epoch 4/5, Batch 7/250, Loss: 1.5276\n",
      "Epoch 4/5, Batch 8/250, Loss: 0.7663\n",
      "Epoch 4/5, Batch 9/250, Loss: 1.8485\n",
      "Epoch 4/5, Batch 10/250, Loss: 0.5682\n",
      "Epoch 4/5, Batch 11/250, Loss: 7.7913\n",
      "Epoch 4/5, Batch 12/250, Loss: 1.4461\n",
      "Epoch 4/5, Batch 13/250, Loss: 0.6082\n",
      "Epoch 4/5, Batch 14/250, Loss: 0.3774\n",
      "Epoch 4/5, Batch 15/250, Loss: 4.5649\n",
      "Epoch 4/5, Batch 16/250, Loss: 3.0295\n",
      "Epoch 4/5, Batch 17/250, Loss: 5.8119\n",
      "Epoch 4/5, Batch 18/250, Loss: 0.9292\n",
      "Epoch 4/5, Batch 19/250, Loss: 3.2867\n",
      "Epoch 4/5, Batch 20/250, Loss: 15.6124\n",
      "Epoch 4/5, Batch 21/250, Loss: 6.4919\n",
      "Epoch 4/5, Batch 22/250, Loss: 0.3859\n",
      "Epoch 4/5, Batch 23/250, Loss: 14.7907\n",
      "Epoch 4/5, Batch 24/250, Loss: 3.0047\n",
      "Epoch 4/5, Batch 25/250, Loss: 7.9338\n",
      "Epoch 4/5, Batch 26/250, Loss: 4.8504\n",
      "Epoch 4/5, Batch 27/250, Loss: 1.7787\n",
      "Epoch 4/5, Batch 28/250, Loss: 4.9162\n",
      "Epoch 4/5, Batch 29/250, Loss: 10.0767\n",
      "Epoch 4/5, Batch 30/250, Loss: 19.9945\n",
      "Epoch 4/5, Batch 31/250, Loss: 2.1813\n",
      "Epoch 4/5, Batch 32/250, Loss: 4.9177\n",
      "Epoch 4/5, Batch 33/250, Loss: 14.0495\n",
      "Epoch 4/5, Batch 34/250, Loss: 18.9147\n",
      "Epoch 4/5, Batch 35/250, Loss: 4.5266\n",
      "Epoch 4/5, Batch 36/250, Loss: 11.5078\n",
      "Epoch 4/5, Batch 37/250, Loss: 3.7264\n",
      "Epoch 4/5, Batch 38/250, Loss: 4.3222\n",
      "Epoch 4/5, Batch 39/250, Loss: 3.7228\n",
      "Epoch 4/5, Batch 40/250, Loss: 5.9843\n",
      "Epoch 4/5, Batch 41/250, Loss: 8.7443\n",
      "Epoch 4/5, Batch 42/250, Loss: 4.1161\n",
      "Epoch 4/5, Batch 43/250, Loss: 10.4276\n",
      "Epoch 4/5, Batch 44/250, Loss: 5.1460\n",
      "Epoch 4/5, Batch 45/250, Loss: 1.3626\n",
      "Epoch 4/5, Batch 46/250, Loss: 18.1474\n",
      "Epoch 4/5, Batch 47/250, Loss: 6.4125\n",
      "Epoch 4/5, Batch 48/250, Loss: 2.5567\n",
      "Epoch 4/5, Batch 49/250, Loss: 12.7853\n",
      "Epoch 4/5, Batch 50/250, Loss: 5.1066\n",
      "Epoch 4/5, Batch 51/250, Loss: 5.8255\n",
      "Epoch 4/5, Batch 52/250, Loss: 4.6989\n",
      "Epoch 4/5, Batch 53/250, Loss: 4.5329\n",
      "Epoch 4/5, Batch 54/250, Loss: 2.5084\n",
      "Epoch 4/5, Batch 55/250, Loss: 8.8398\n",
      "Epoch 4/5, Batch 56/250, Loss: 0.9936\n",
      "Epoch 4/5, Batch 57/250, Loss: 3.9564\n",
      "Epoch 4/5, Batch 58/250, Loss: 4.0984\n",
      "Epoch 4/5, Batch 59/250, Loss: 1.1911\n",
      "Epoch 4/5, Batch 60/250, Loss: 1.4975\n",
      "Epoch 4/5, Batch 61/250, Loss: 2.3273\n",
      "Epoch 4/5, Batch 62/250, Loss: 7.9987\n",
      "Epoch 4/5, Batch 63/250, Loss: 7.6349\n",
      "Epoch 4/5, Batch 64/250, Loss: 7.6816\n",
      "Epoch 4/5, Batch 65/250, Loss: 0.3748\n",
      "Epoch 4/5, Batch 66/250, Loss: 1.9390\n",
      "Epoch 4/5, Batch 67/250, Loss: 1.7146\n",
      "Epoch 4/5, Batch 68/250, Loss: 4.2576\n",
      "Epoch 4/5, Batch 69/250, Loss: 1.2955\n",
      "Epoch 4/5, Batch 70/250, Loss: 2.6918\n",
      "Epoch 4/5, Batch 71/250, Loss: 4.2869\n",
      "Epoch 4/5, Batch 72/250, Loss: 10.4936\n",
      "Epoch 4/5, Batch 73/250, Loss: 0.4287\n",
      "Epoch 4/5, Batch 74/250, Loss: 1.2394\n",
      "Epoch 4/5, Batch 75/250, Loss: 3.0352\n",
      "Epoch 4/5, Batch 76/250, Loss: 2.3496\n",
      "Epoch 4/5, Batch 77/250, Loss: 2.2843\n",
      "Epoch 4/5, Batch 78/250, Loss: 4.1423\n",
      "Epoch 4/5, Batch 79/250, Loss: 1.8741\n",
      "Epoch 4/5, Batch 80/250, Loss: 3.7775\n",
      "Epoch 4/5, Batch 81/250, Loss: 5.0781\n",
      "Epoch 4/5, Batch 82/250, Loss: 0.6908\n",
      "Epoch 4/5, Batch 83/250, Loss: 4.5286\n",
      "Epoch 4/5, Batch 84/250, Loss: 3.9037\n",
      "Epoch 4/5, Batch 85/250, Loss: 8.0695\n",
      "Epoch 4/5, Batch 86/250, Loss: 4.4993\n",
      "Epoch 4/5, Batch 87/250, Loss: 2.6301\n",
      "Epoch 4/5, Batch 88/250, Loss: 3.9609\n",
      "Epoch 4/5, Batch 89/250, Loss: 1.7159\n",
      "Epoch 4/5, Batch 90/250, Loss: 5.7477\n",
      "Epoch 4/5, Batch 91/250, Loss: 4.5451\n",
      "Epoch 4/5, Batch 92/250, Loss: 5.0174\n",
      "Epoch 4/5, Batch 93/250, Loss: 2.1220\n",
      "Epoch 4/5, Batch 94/250, Loss: 3.6913\n",
      "Epoch 4/5, Batch 95/250, Loss: 8.3740\n",
      "Epoch 4/5, Batch 96/250, Loss: 9.5951\n",
      "Epoch 4/5, Batch 97/250, Loss: 2.1303\n",
      "Epoch 4/5, Batch 98/250, Loss: 10.1993\n",
      "Epoch 4/5, Batch 99/250, Loss: 4.5297\n",
      "Epoch 4/5, Batch 100/250, Loss: 3.3618\n",
      "Epoch 4/5, Batch 101/250, Loss: 1.6573\n",
      "Epoch 4/5, Batch 102/250, Loss: 0.6505\n",
      "Epoch 4/5, Batch 103/250, Loss: 11.7234\n",
      "Epoch 4/5, Batch 104/250, Loss: 3.5868\n",
      "Epoch 4/5, Batch 105/250, Loss: 9.5382\n",
      "Epoch 4/5, Batch 106/250, Loss: 1.4466\n",
      "Epoch 4/5, Batch 107/250, Loss: 21.9664\n",
      "Epoch 4/5, Batch 108/250, Loss: 3.8280\n",
      "Epoch 4/5, Batch 109/250, Loss: 7.5908\n",
      "Epoch 4/5, Batch 110/250, Loss: 2.0082\n",
      "Epoch 4/5, Batch 111/250, Loss: 0.9609\n",
      "Epoch 4/5, Batch 112/250, Loss: 2.3329\n",
      "Epoch 4/5, Batch 113/250, Loss: 6.1836\n",
      "Epoch 4/5, Batch 114/250, Loss: 13.0899\n",
      "Epoch 4/5, Batch 115/250, Loss: 4.0403\n",
      "Epoch 4/5, Batch 116/250, Loss: 8.5000\n",
      "Epoch 4/5, Batch 117/250, Loss: 1.7324\n",
      "Epoch 4/5, Batch 118/250, Loss: 2.0883\n",
      "Epoch 4/5, Batch 119/250, Loss: 5.8437\n",
      "Epoch 4/5, Batch 120/250, Loss: 9.0128\n",
      "Epoch 4/5, Batch 121/250, Loss: 10.6583\n",
      "Epoch 4/5, Batch 122/250, Loss: 16.9510\n",
      "Epoch 4/5, Batch 123/250, Loss: 4.8444\n",
      "Epoch 4/5, Batch 124/250, Loss: 7.4338\n",
      "Epoch 4/5, Batch 125/250, Loss: 6.1539\n",
      "Epoch 4/5, Batch 126/250, Loss: 6.0399\n",
      "Epoch 4/5, Batch 127/250, Loss: 6.3399\n",
      "Epoch 4/5, Batch 128/250, Loss: 4.3120\n",
      "Epoch 4/5, Batch 129/250, Loss: 6.5545\n",
      "Epoch 4/5, Batch 130/250, Loss: 6.7019\n",
      "Epoch 4/5, Batch 131/250, Loss: 5.0336\n",
      "Epoch 4/5, Batch 132/250, Loss: 4.4303\n",
      "Epoch 4/5, Batch 133/250, Loss: 3.7845\n",
      "Epoch 4/5, Batch 134/250, Loss: 3.1934\n",
      "Epoch 4/5, Batch 135/250, Loss: 4.3811\n",
      "Epoch 4/5, Batch 136/250, Loss: 5.5477\n",
      "Epoch 4/5, Batch 137/250, Loss: 1.3203\n",
      "Epoch 4/5, Batch 138/250, Loss: 7.2183\n",
      "Epoch 4/5, Batch 139/250, Loss: 1.3751\n",
      "Epoch 4/5, Batch 140/250, Loss: 4.4067\n",
      "Epoch 4/5, Batch 141/250, Loss: 0.5114\n",
      "Epoch 4/5, Batch 142/250, Loss: 1.9084\n",
      "Epoch 4/5, Batch 143/250, Loss: 4.3647\n",
      "Epoch 4/5, Batch 144/250, Loss: 5.3927\n",
      "Epoch 4/5, Batch 145/250, Loss: 3.1893\n",
      "Epoch 4/5, Batch 146/250, Loss: 3.2189\n",
      "Epoch 4/5, Batch 147/250, Loss: 3.6551\n",
      "Epoch 4/5, Batch 148/250, Loss: 4.4782\n",
      "Epoch 4/5, Batch 149/250, Loss: 0.5156\n",
      "Epoch 4/5, Batch 150/250, Loss: 11.2738\n",
      "Epoch 4/5, Batch 151/250, Loss: 6.1334\n",
      "Epoch 4/5, Batch 152/250, Loss: 2.1103\n",
      "Epoch 4/5, Batch 153/250, Loss: 2.5887\n",
      "Epoch 4/5, Batch 154/250, Loss: 13.9219\n",
      "Epoch 4/5, Batch 155/250, Loss: 1.2327\n",
      "Epoch 4/5, Batch 156/250, Loss: 6.0665\n",
      "Epoch 4/5, Batch 157/250, Loss: 4.1774\n",
      "Epoch 4/5, Batch 158/250, Loss: 5.8423\n",
      "Epoch 4/5, Batch 159/250, Loss: 4.8431\n",
      "Epoch 4/5, Batch 160/250, Loss: 2.6298\n",
      "Epoch 4/5, Batch 161/250, Loss: 5.3149\n",
      "Epoch 4/5, Batch 162/250, Loss: 5.8217\n",
      "Epoch 4/5, Batch 163/250, Loss: 9.4460\n",
      "Epoch 4/5, Batch 164/250, Loss: 6.3274\n",
      "Epoch 4/5, Batch 165/250, Loss: 8.0311\n",
      "Epoch 4/5, Batch 166/250, Loss: 9.9903\n",
      "Epoch 4/5, Batch 167/250, Loss: 10.0357\n",
      "Epoch 4/5, Batch 168/250, Loss: 10.0936\n",
      "Epoch 4/5, Batch 169/250, Loss: 8.0376\n",
      "Epoch 4/5, Batch 170/250, Loss: 3.3759\n",
      "Epoch 4/5, Batch 171/250, Loss: 4.2132\n",
      "Epoch 4/5, Batch 172/250, Loss: 8.3460\n",
      "Epoch 4/5, Batch 173/250, Loss: 8.2656\n",
      "Epoch 4/5, Batch 174/250, Loss: 3.9264\n",
      "Epoch 4/5, Batch 175/250, Loss: 10.7971\n",
      "Epoch 4/5, Batch 176/250, Loss: 5.9623\n",
      "Epoch 4/5, Batch 177/250, Loss: 4.8932\n",
      "Epoch 4/5, Batch 178/250, Loss: 5.8199\n",
      "Epoch 4/5, Batch 179/250, Loss: 4.2493\n",
      "Epoch 4/5, Batch 180/250, Loss: 1.7011\n",
      "Epoch 4/5, Batch 181/250, Loss: 7.1682\n",
      "Epoch 4/5, Batch 182/250, Loss: 5.8160\n",
      "Epoch 4/5, Batch 183/250, Loss: 14.3021\n",
      "Epoch 4/5, Batch 184/250, Loss: 3.3905\n",
      "Epoch 4/5, Batch 185/250, Loss: 7.2200\n",
      "Epoch 4/5, Batch 186/250, Loss: 8.3062\n",
      "Epoch 4/5, Batch 187/250, Loss: 4.8160\n",
      "Epoch 4/5, Batch 188/250, Loss: 6.7586\n",
      "Epoch 4/5, Batch 189/250, Loss: 1.4189\n",
      "Epoch 4/5, Batch 190/250, Loss: 12.2102\n",
      "Epoch 4/5, Batch 191/250, Loss: 8.1760\n",
      "Epoch 4/5, Batch 192/250, Loss: 6.6495\n",
      "Epoch 4/5, Batch 193/250, Loss: 3.0494\n",
      "Epoch 4/5, Batch 194/250, Loss: 5.0918\n",
      "Epoch 4/5, Batch 195/250, Loss: 4.2895\n",
      "Epoch 4/5, Batch 196/250, Loss: 5.8924\n",
      "Epoch 4/5, Batch 197/250, Loss: 7.0704\n",
      "Epoch 4/5, Batch 198/250, Loss: 3.3951\n",
      "Epoch 4/5, Batch 199/250, Loss: 13.8143\n",
      "Epoch 4/5, Batch 200/250, Loss: 1.4967\n",
      "Epoch 4/5, Batch 201/250, Loss: 6.8083\n",
      "Epoch 4/5, Batch 202/250, Loss: 3.9606\n",
      "Epoch 4/5, Batch 203/250, Loss: 3.8898\n",
      "Epoch 4/5, Batch 204/250, Loss: 6.7515\n",
      "Epoch 4/5, Batch 205/250, Loss: 2.4759\n",
      "Epoch 4/5, Batch 206/250, Loss: 3.5801\n",
      "Epoch 4/5, Batch 207/250, Loss: 12.5209\n",
      "Epoch 4/5, Batch 208/250, Loss: 8.1714\n",
      "Epoch 4/5, Batch 209/250, Loss: 2.8303\n",
      "Epoch 4/5, Batch 210/250, Loss: 13.8828\n",
      "Epoch 4/5, Batch 211/250, Loss: 3.4623\n",
      "Epoch 4/5, Batch 212/250, Loss: 9.4494\n",
      "Epoch 4/5, Batch 213/250, Loss: 4.5400\n",
      "Epoch 4/5, Batch 214/250, Loss: 3.7154\n",
      "Epoch 4/5, Batch 215/250, Loss: 9.7046\n",
      "Epoch 4/5, Batch 216/250, Loss: 10.9392\n",
      "Epoch 4/5, Batch 217/250, Loss: 9.4046\n",
      "Epoch 4/5, Batch 218/250, Loss: 3.2104\n",
      "Epoch 4/5, Batch 219/250, Loss: 14.6230\n",
      "Epoch 4/5, Batch 220/250, Loss: 8.7301\n",
      "Epoch 4/5, Batch 221/250, Loss: 10.6945\n",
      "Epoch 4/5, Batch 222/250, Loss: 2.9746\n",
      "Epoch 4/5, Batch 223/250, Loss: 16.7862\n",
      "Epoch 4/5, Batch 224/250, Loss: 13.9855\n",
      "Epoch 4/5, Batch 225/250, Loss: 2.8251\n",
      "Epoch 4/5, Batch 226/250, Loss: 7.7716\n",
      "Epoch 4/5, Batch 227/250, Loss: 2.0053\n",
      "Epoch 4/5, Batch 228/250, Loss: 1.3232\n",
      "Epoch 4/5, Batch 229/250, Loss: 3.9284\n",
      "Epoch 4/5, Batch 230/250, Loss: 2.3975\n",
      "Epoch 4/5, Batch 231/250, Loss: 7.6000\n",
      "Epoch 4/5, Batch 232/250, Loss: 4.4356\n",
      "Epoch 4/5, Batch 233/250, Loss: 6.8496\n",
      "Epoch 4/5, Batch 234/250, Loss: 3.6009\n",
      "Epoch 4/5, Batch 235/250, Loss: 26.5653\n",
      "Epoch 4/5, Batch 236/250, Loss: 15.0539\n",
      "Epoch 4/5, Batch 237/250, Loss: 10.2927\n",
      "Epoch 4/5, Batch 238/250, Loss: 2.6197\n",
      "Epoch 4/5, Batch 239/250, Loss: 12.1502\n",
      "Epoch 4/5, Batch 240/250, Loss: 5.6386\n",
      "Epoch 4/5, Batch 241/250, Loss: 8.9439\n",
      "Epoch 4/5, Batch 242/250, Loss: 6.9699\n",
      "Epoch 4/5, Batch 243/250, Loss: 1.5500\n",
      "Epoch 4/5, Batch 244/250, Loss: 6.9484\n",
      "Epoch 4/5, Batch 245/250, Loss: 3.7967\n",
      "Epoch 4/5, Batch 246/250, Loss: 9.7679\n",
      "Epoch 4/5, Batch 247/250, Loss: 8.4642\n",
      "Epoch 4/5, Batch 248/250, Loss: 9.7152\n",
      "Epoch 4/5, Batch 249/250, Loss: 9.6310\n",
      "Epoch 4/5, Batch 250/250, Loss: 9.2052\n",
      "Epoch 4/5\n",
      "Training Loss: 5.8529\n",
      "Validation Loss: 51.9823\n",
      "Validation Accuracy: 0.8654\n",
      "\n",
      "Epoch 5/5, Batch 1/250, Loss: 2.0589\n",
      "Epoch 5/5, Batch 2/250, Loss: 6.8668\n",
      "Epoch 5/5, Batch 3/250, Loss: 4.0630\n",
      "Epoch 5/5, Batch 4/250, Loss: 4.7702\n",
      "Epoch 5/5, Batch 5/250, Loss: 5.7410\n",
      "Epoch 5/5, Batch 6/250, Loss: 6.9474\n",
      "Epoch 5/5, Batch 7/250, Loss: 4.3112\n",
      "Epoch 5/5, Batch 8/250, Loss: 3.4917\n",
      "Epoch 5/5, Batch 9/250, Loss: 5.1017\n",
      "Epoch 5/5, Batch 10/250, Loss: 2.7043\n",
      "Epoch 5/5, Batch 11/250, Loss: 6.0088\n",
      "Epoch 5/5, Batch 12/250, Loss: 1.1341\n",
      "Epoch 5/5, Batch 13/250, Loss: 2.0337\n",
      "Epoch 5/5, Batch 14/250, Loss: 1.7409\n",
      "Epoch 5/5, Batch 15/250, Loss: 3.0117\n",
      "Epoch 5/5, Batch 16/250, Loss: 0.7934\n",
      "Epoch 5/5, Batch 17/250, Loss: 2.1334\n",
      "Epoch 5/5, Batch 18/250, Loss: 8.4878\n",
      "Epoch 5/5, Batch 19/250, Loss: 1.0964\n",
      "Epoch 5/5, Batch 20/250, Loss: 2.9993\n",
      "Epoch 5/5, Batch 21/250, Loss: 3.2286\n",
      "Epoch 5/5, Batch 22/250, Loss: 5.0986\n",
      "Epoch 5/5, Batch 23/250, Loss: 0.3215\n",
      "Epoch 5/5, Batch 24/250, Loss: 2.4369\n",
      "Epoch 5/5, Batch 25/250, Loss: 0.4875\n",
      "Epoch 5/5, Batch 26/250, Loss: 0.5955\n",
      "Epoch 5/5, Batch 27/250, Loss: 0.6298\n",
      "Epoch 5/5, Batch 28/250, Loss: 1.0794\n",
      "Epoch 5/5, Batch 29/250, Loss: 8.0385\n",
      "Epoch 5/5, Batch 30/250, Loss: 1.4503\n",
      "Epoch 5/5, Batch 31/250, Loss: 3.1663\n",
      "Epoch 5/5, Batch 32/250, Loss: 0.3652\n",
      "Epoch 5/5, Batch 33/250, Loss: 6.9401\n",
      "Epoch 5/5, Batch 34/250, Loss: 1.9826\n",
      "Epoch 5/5, Batch 35/250, Loss: 1.2586\n",
      "Epoch 5/5, Batch 36/250, Loss: 2.9380\n",
      "Epoch 5/5, Batch 37/250, Loss: 0.3255\n",
      "Epoch 5/5, Batch 38/250, Loss: 5.0663\n",
      "Epoch 5/5, Batch 39/250, Loss: 3.2582\n",
      "Epoch 5/5, Batch 40/250, Loss: 1.5247\n",
      "Epoch 5/5, Batch 41/250, Loss: 5.8285\n",
      "Epoch 5/5, Batch 42/250, Loss: 0.2423\n",
      "Epoch 5/5, Batch 43/250, Loss: 2.0122\n",
      "Epoch 5/5, Batch 44/250, Loss: 6.1942\n",
      "Epoch 5/5, Batch 45/250, Loss: 4.0859\n",
      "Epoch 5/5, Batch 46/250, Loss: 0.5321\n",
      "Epoch 5/5, Batch 47/250, Loss: 2.6876\n",
      "Epoch 5/5, Batch 48/250, Loss: 1.5215\n",
      "Epoch 5/5, Batch 49/250, Loss: 4.3362\n",
      "Epoch 5/5, Batch 50/250, Loss: 0.8964\n",
      "Epoch 5/5, Batch 51/250, Loss: 1.0464\n",
      "Epoch 5/5, Batch 52/250, Loss: 1.5005\n",
      "Epoch 5/5, Batch 53/250, Loss: 2.1381\n",
      "Epoch 5/5, Batch 54/250, Loss: 4.7149\n",
      "Epoch 5/5, Batch 55/250, Loss: 0.5757\n",
      "Epoch 5/5, Batch 56/250, Loss: 4.5846\n",
      "Epoch 5/5, Batch 57/250, Loss: 2.5990\n",
      "Epoch 5/5, Batch 58/250, Loss: 0.8853\n",
      "Epoch 5/5, Batch 59/250, Loss: 3.4540\n",
      "Epoch 5/5, Batch 60/250, Loss: 4.4571\n",
      "Epoch 5/5, Batch 61/250, Loss: 8.3488\n",
      "Epoch 5/5, Batch 62/250, Loss: 7.8045\n",
      "Epoch 5/5, Batch 63/250, Loss: 0.9219\n",
      "Epoch 5/5, Batch 64/250, Loss: 2.6040\n",
      "Epoch 5/5, Batch 65/250, Loss: 1.0740\n",
      "Epoch 5/5, Batch 66/250, Loss: 0.7926\n",
      "Epoch 5/5, Batch 67/250, Loss: 0.6343\n",
      "Epoch 5/5, Batch 68/250, Loss: 0.9476\n",
      "Epoch 5/5, Batch 69/250, Loss: 6.4241\n",
      "Epoch 5/5, Batch 70/250, Loss: 18.0022\n",
      "Epoch 5/5, Batch 71/250, Loss: 1.5361\n",
      "Epoch 5/5, Batch 72/250, Loss: 2.1828\n",
      "Epoch 5/5, Batch 73/250, Loss: 6.2639\n",
      "Epoch 5/5, Batch 74/250, Loss: 7.0476\n",
      "Epoch 5/5, Batch 75/250, Loss: 7.8937\n",
      "Epoch 5/5, Batch 76/250, Loss: 2.1351\n",
      "Epoch 5/5, Batch 77/250, Loss: 2.5762\n",
      "Epoch 5/5, Batch 78/250, Loss: 0.6672\n",
      "Epoch 5/5, Batch 79/250, Loss: 4.7325\n",
      "Epoch 5/5, Batch 80/250, Loss: 6.9700\n",
      "Epoch 5/5, Batch 81/250, Loss: 0.8122\n",
      "Epoch 5/5, Batch 82/250, Loss: 4.5477\n",
      "Epoch 5/5, Batch 83/250, Loss: 1.6513\n",
      "Epoch 5/5, Batch 84/250, Loss: 3.7598\n",
      "Epoch 5/5, Batch 85/250, Loss: 2.7247\n",
      "Epoch 5/5, Batch 86/250, Loss: 3.0282\n",
      "Epoch 5/5, Batch 87/250, Loss: 1.7578\n",
      "Epoch 5/5, Batch 88/250, Loss: 4.9029\n",
      "Epoch 5/5, Batch 89/250, Loss: 3.7522\n",
      "Epoch 5/5, Batch 90/250, Loss: 3.6060\n",
      "Epoch 5/5, Batch 91/250, Loss: 2.1816\n",
      "Epoch 5/5, Batch 92/250, Loss: 5.8094\n",
      "Epoch 5/5, Batch 93/250, Loss: 1.9649\n",
      "Epoch 5/5, Batch 94/250, Loss: 5.0543\n",
      "Epoch 5/5, Batch 95/250, Loss: 0.7191\n",
      "Epoch 5/5, Batch 96/250, Loss: 0.9621\n",
      "Epoch 5/5, Batch 97/250, Loss: 5.3160\n",
      "Epoch 5/5, Batch 98/250, Loss: 5.4487\n",
      "Epoch 5/5, Batch 99/250, Loss: 0.9328\n",
      "Epoch 5/5, Batch 100/250, Loss: 0.2094\n",
      "Epoch 5/5, Batch 101/250, Loss: 1.1109\n",
      "Epoch 5/5, Batch 102/250, Loss: 4.4263\n",
      "Epoch 5/5, Batch 103/250, Loss: 4.5438\n",
      "Epoch 5/5, Batch 104/250, Loss: 0.6943\n",
      "Epoch 5/5, Batch 105/250, Loss: 2.3089\n",
      "Epoch 5/5, Batch 106/250, Loss: 0.5328\n",
      "Epoch 5/5, Batch 107/250, Loss: 1.2508\n",
      "Epoch 5/5, Batch 108/250, Loss: 8.7767\n",
      "Epoch 5/5, Batch 109/250, Loss: 0.2164\n",
      "Epoch 5/5, Batch 110/250, Loss: 1.8607\n",
      "Epoch 5/5, Batch 111/250, Loss: 0.9464\n",
      "Epoch 5/5, Batch 112/250, Loss: 1.2272\n",
      "Epoch 5/5, Batch 113/250, Loss: 1.3011\n",
      "Epoch 5/5, Batch 114/250, Loss: 0.6310\n",
      "Epoch 5/5, Batch 115/250, Loss: 1.8155\n",
      "Epoch 5/5, Batch 116/250, Loss: 2.9912\n",
      "Epoch 5/5, Batch 117/250, Loss: 9.8994\n",
      "Epoch 5/5, Batch 118/250, Loss: 12.4802\n",
      "Epoch 5/5, Batch 119/250, Loss: 1.0960\n",
      "Epoch 5/5, Batch 120/250, Loss: 9.8578\n",
      "Epoch 5/5, Batch 121/250, Loss: 6.1038\n",
      "Epoch 5/5, Batch 122/250, Loss: 1.8553\n",
      "Epoch 5/5, Batch 123/250, Loss: 11.5867\n",
      "Epoch 5/5, Batch 124/250, Loss: 0.5095\n",
      "Epoch 5/5, Batch 125/250, Loss: 0.3138\n",
      "Epoch 5/5, Batch 126/250, Loss: 6.8195\n",
      "Epoch 5/5, Batch 127/250, Loss: 20.8156\n",
      "Epoch 5/5, Batch 128/250, Loss: 1.4581\n",
      "Epoch 5/5, Batch 129/250, Loss: 5.7182\n",
      "Epoch 5/5, Batch 130/250, Loss: 0.9659\n",
      "Epoch 5/5, Batch 131/250, Loss: 8.3191\n",
      "Epoch 5/5, Batch 132/250, Loss: 6.1712\n",
      "Epoch 5/5, Batch 133/250, Loss: 4.3136\n",
      "Epoch 5/5, Batch 134/250, Loss: 1.2033\n",
      "Epoch 5/5, Batch 135/250, Loss: 8.4980\n",
      "Epoch 5/5, Batch 136/250, Loss: 6.0747\n",
      "Epoch 5/5, Batch 137/250, Loss: 1.9722\n",
      "Epoch 5/5, Batch 138/250, Loss: 0.6078\n",
      "Epoch 5/5, Batch 139/250, Loss: 3.0150\n",
      "Epoch 5/5, Batch 140/250, Loss: 2.9666\n",
      "Epoch 5/5, Batch 141/250, Loss: 5.1563\n",
      "Epoch 5/5, Batch 142/250, Loss: 6.9768\n",
      "Epoch 5/5, Batch 143/250, Loss: 2.5777\n",
      "Epoch 5/5, Batch 144/250, Loss: 1.1896\n",
      "Epoch 5/5, Batch 145/250, Loss: 3.2218\n",
      "Epoch 5/5, Batch 146/250, Loss: 5.6986\n",
      "Epoch 5/5, Batch 147/250, Loss: 4.5516\n",
      "Epoch 5/5, Batch 148/250, Loss: 0.9628\n",
      "Epoch 5/5, Batch 149/250, Loss: 3.9674\n",
      "Epoch 5/5, Batch 150/250, Loss: 0.8896\n",
      "Epoch 5/5, Batch 151/250, Loss: 3.4864\n",
      "Epoch 5/5, Batch 152/250, Loss: 18.9312\n",
      "Epoch 5/5, Batch 153/250, Loss: 3.0913\n",
      "Epoch 5/5, Batch 154/250, Loss: 5.1823\n",
      "Epoch 5/5, Batch 155/250, Loss: 0.6154\n",
      "Epoch 5/5, Batch 156/250, Loss: 2.5158\n",
      "Epoch 5/5, Batch 157/250, Loss: 0.8360\n",
      "Epoch 5/5, Batch 158/250, Loss: 3.6891\n",
      "Epoch 5/5, Batch 159/250, Loss: 1.0048\n",
      "Epoch 5/5, Batch 160/250, Loss: 3.9434\n",
      "Epoch 5/5, Batch 161/250, Loss: 4.8366\n",
      "Epoch 5/5, Batch 162/250, Loss: 2.2449\n",
      "Epoch 5/5, Batch 163/250, Loss: 5.0347\n",
      "Epoch 5/5, Batch 164/250, Loss: 6.7580\n",
      "Epoch 5/5, Batch 165/250, Loss: 7.9733\n",
      "Epoch 5/5, Batch 166/250, Loss: 0.8092\n",
      "Epoch 5/5, Batch 167/250, Loss: 2.7738\n",
      "Epoch 5/5, Batch 168/250, Loss: 4.6518\n",
      "Epoch 5/5, Batch 169/250, Loss: 10.4755\n",
      "Epoch 5/5, Batch 170/250, Loss: 6.8385\n",
      "Epoch 5/5, Batch 171/250, Loss: 9.9897\n",
      "Epoch 5/5, Batch 172/250, Loss: 7.9277\n",
      "Epoch 5/5, Batch 173/250, Loss: 20.5696\n",
      "Epoch 5/5, Batch 174/250, Loss: 8.5674\n",
      "Epoch 5/5, Batch 175/250, Loss: 1.6320\n",
      "Epoch 5/5, Batch 176/250, Loss: 3.3212\n",
      "Epoch 5/5, Batch 177/250, Loss: 7.4233\n",
      "Epoch 5/5, Batch 178/250, Loss: 5.2954\n",
      "Epoch 5/5, Batch 179/250, Loss: 4.2429\n",
      "Epoch 5/5, Batch 180/250, Loss: 6.5355\n",
      "Epoch 5/5, Batch 181/250, Loss: 2.8654\n",
      "Epoch 5/5, Batch 182/250, Loss: 1.2861\n",
      "Epoch 5/5, Batch 183/250, Loss: 5.1808\n",
      "Epoch 5/5, Batch 184/250, Loss: 2.6539\n",
      "Epoch 5/5, Batch 185/250, Loss: 7.1155\n",
      "Epoch 5/5, Batch 186/250, Loss: 2.6003\n",
      "Epoch 5/5, Batch 187/250, Loss: 4.6761\n",
      "Epoch 5/5, Batch 188/250, Loss: 6.0197\n",
      "Epoch 5/5, Batch 189/250, Loss: 1.0588\n",
      "Epoch 5/5, Batch 190/250, Loss: 3.0204\n",
      "Epoch 5/5, Batch 191/250, Loss: 2.3489\n",
      "Epoch 5/5, Batch 192/250, Loss: 17.6295\n",
      "Epoch 5/5, Batch 193/250, Loss: 0.7995\n",
      "Epoch 5/5, Batch 194/250, Loss: 1.2080\n",
      "Epoch 5/5, Batch 195/250, Loss: 7.5554\n",
      "Epoch 5/5, Batch 196/250, Loss: 8.5371\n",
      "Epoch 5/5, Batch 197/250, Loss: 1.5292\n",
      "Epoch 5/5, Batch 198/250, Loss: 5.6261\n",
      "Epoch 5/5, Batch 199/250, Loss: 1.7586\n",
      "Epoch 5/5, Batch 200/250, Loss: 5.0313\n",
      "Epoch 5/5, Batch 201/250, Loss: 2.7350\n",
      "Epoch 5/5, Batch 202/250, Loss: 1.4818\n",
      "Epoch 5/5, Batch 203/250, Loss: 9.0578\n",
      "Epoch 5/5, Batch 204/250, Loss: 2.0091\n",
      "Epoch 5/5, Batch 205/250, Loss: 1.3868\n",
      "Epoch 5/5, Batch 206/250, Loss: 3.3734\n",
      "Epoch 5/5, Batch 207/250, Loss: 0.9506\n",
      "Epoch 5/5, Batch 208/250, Loss: 11.3662\n",
      "Epoch 5/5, Batch 209/250, Loss: 0.5779\n",
      "Epoch 5/5, Batch 210/250, Loss: 7.8673\n",
      "Epoch 5/5, Batch 211/250, Loss: 7.9595\n",
      "Epoch 5/5, Batch 212/250, Loss: 0.9215\n",
      "Epoch 5/5, Batch 213/250, Loss: 1.1013\n",
      "Epoch 5/5, Batch 214/250, Loss: 4.0957\n",
      "Epoch 5/5, Batch 215/250, Loss: 3.0244\n",
      "Epoch 5/5, Batch 216/250, Loss: 0.5122\n",
      "Epoch 5/5, Batch 217/250, Loss: 3.5573\n",
      "Epoch 5/5, Batch 218/250, Loss: 0.4467\n",
      "Epoch 5/5, Batch 219/250, Loss: 3.4120\n",
      "Epoch 5/5, Batch 220/250, Loss: 1.0150\n",
      "Epoch 5/5, Batch 221/250, Loss: 1.8129\n",
      "Epoch 5/5, Batch 222/250, Loss: 6.2615\n",
      "Epoch 5/5, Batch 223/250, Loss: 0.5671\n",
      "Epoch 5/5, Batch 224/250, Loss: 3.2629\n",
      "Epoch 5/5, Batch 225/250, Loss: 3.6881\n",
      "Epoch 5/5, Batch 226/250, Loss: 4.7442\n",
      "Epoch 5/5, Batch 227/250, Loss: 2.4679\n",
      "Epoch 5/5, Batch 228/250, Loss: 1.3743\n",
      "Epoch 5/5, Batch 229/250, Loss: 1.2387\n",
      "Epoch 5/5, Batch 230/250, Loss: 2.9679\n",
      "Epoch 5/5, Batch 231/250, Loss: 1.0614\n",
      "Epoch 5/5, Batch 232/250, Loss: 2.2967\n",
      "Epoch 5/5, Batch 233/250, Loss: 1.9169\n",
      "Epoch 5/5, Batch 234/250, Loss: 7.1940\n",
      "Epoch 5/5, Batch 235/250, Loss: 2.9227\n",
      "Epoch 5/5, Batch 236/250, Loss: 0.4567\n",
      "Epoch 5/5, Batch 237/250, Loss: 3.8095\n",
      "Epoch 5/5, Batch 238/250, Loss: 1.6852\n",
      "Epoch 5/5, Batch 239/250, Loss: 8.2418\n",
      "Epoch 5/5, Batch 240/250, Loss: 5.5066\n",
      "Epoch 5/5, Batch 241/250, Loss: 4.8184\n",
      "Epoch 5/5, Batch 242/250, Loss: 7.8397\n",
      "Epoch 5/5, Batch 243/250, Loss: 8.0180\n",
      "Epoch 5/5, Batch 244/250, Loss: 6.8132\n",
      "Epoch 5/5, Batch 245/250, Loss: 7.7401\n",
      "Epoch 5/5, Batch 246/250, Loss: 15.5938\n",
      "Epoch 5/5, Batch 247/250, Loss: 12.2262\n",
      "Epoch 5/5, Batch 248/250, Loss: 8.9145\n",
      "Epoch 5/5, Batch 249/250, Loss: 2.3731\n",
      "Epoch 5/5, Batch 250/250, Loss: 2.9222\n",
      "Epoch 5/5\n",
      "Training Loss: 3.9937\n",
      "Validation Loss: 42.4656\n",
      "Validation Accuracy: 0.8663\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Training loop\n",
    "for epoch in range(epochs):\n",
    "    model.train()\n",
    "    total_loss = 0\n",
    "    for i, batch in enumerate(train_loader):\n",
    "        optimizer.zero_grad()\n",
    "        input_ids = batch['input_ids'].to(device)\n",
    "        attention_mask = batch['attention_mask'].to(device)\n",
    "        labels = batch['labels'].unsqueeze(1).float().to(device)  # or labels = batch['labels'].view(-1,1).float().to(device)\n",
    "        outputs = model(input_ids, attention_mask=attention_mask, labels=labels)\n",
    "        logits = outputs.logits\n",
    "        loss = loss_fn(logits, labels).sum()\n",
    "        print(f'Epoch {epoch+1}/{epochs}, Batch {i+1}/{len(train_loader)}, Loss: {loss.item():.4f}')\n",
    "        total_loss += loss.item()\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "    \n",
    "    average_train_loss = total_loss/len(train_loader)\n",
    "\n",
    "    print(f'Epoch {epoch+1}/{epochs}')\n",
    "    print(f'Training Loss: {average_train_loss:.4f}')\n",
    "\n",
    "    # Validation\n",
    "    model.eval()\n",
    "    test_loss = 0\n",
    "    correct = 0\n",
    "    total = 0\n",
    "    with torch.no_grad():\n",
    "        for batch in test_loader:\n",
    "            input_ids = batch['input_ids'].to(device)\n",
    "            attention_mask = batch['attention_mask'].to(device)\n",
    "            labels = batch['labels'].unsqueeze(1).float().to(device)  # or labels = batch['labels'].view(-1,1).float().to(device)\n",
    "            outputs = model(input_ids, attention_mask=attention_mask)\n",
    "            logits = outputs.logits\n",
    "            loss = loss_fn(logits, labels).sum()\n",
    "            test_loss += loss.item()\n",
    "\n",
    "            # Compute binary predictions\n",
    "            predicted = (torch.sigmoid(logits)>0.5).float()\n",
    "            total += labels.size(0)\n",
    "            correct += (predicted==labels).sum().item()\n",
    "    \n",
    "    average_test_loss = test_loss/len(test_loader)\n",
    "    test_accuracy = correct/total\n",
    "\n",
    "    print(f'Validation Loss: {average_test_loss:.4f}')\n",
    "    print(f'Validation Accuracy: {test_accuracy:.4f}\\n')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = df_train_sentiment[\"description\"]\n",
    "y_train = df_train_sentiment[\"sentiment\"]\n",
    "X_test = df_test_sentiment[\"description\"]\n",
    "y_test = df_test_sentiment[\"sentiment\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "def predict_sentiment(X):\n",
    "    res = []\n",
    "    for text in X:\n",
    "        # Tokenize input text\n",
    "        inputs = tokenizer(text, padding='max_length', truncation=True, max_length=max_length, return_tensors='pt').to(device)\n",
    "        # Use model for inference\n",
    "        outputs = model(**inputs)\n",
    "        # Get predicted label (1 or 0)\n",
    "        predicted_label = (torch.sigmoid(outputs.logits)>0.5).float().item()\n",
    "        res.append(predicted_label)\n",
    "    return res"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_train_pred = predict_sentiment(X_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_test_pred = predict_sentiment(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import classification_report"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "         0.0       1.00      0.98      0.99     12500\n",
      "         1.0       0.98      1.00      0.99     12500\n",
      "\n",
      "    accuracy                           0.99     25000\n",
      "   macro avg       0.99      0.99      0.99     25000\n",
      "weighted avg       0.99      0.99      0.99     25000\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_train, y_train_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "         0.0       0.91      0.81      0.86     12500\n",
      "         1.0       0.83      0.92      0.87     12500\n",
      "\n",
      "    accuracy                           0.87     25000\n",
      "   macro avg       0.87      0.87      0.87     25000\n",
      "weighted avg       0.87      0.87      0.87     25000\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(classification_report(y_test, y_test_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define function to perform sentiment analysis\n",
    "def get_sentiment(text):\n",
    "    # Tokenize input text\n",
    "    inputs = tokenizer(text, padding='max_length', truncation=True, max_length=max_length, return_tensors='pt').to(device)\n",
    "    # Use model for inference\n",
    "    outputs = model(**inputs)\n",
    "    # Get predicted label (positive or negative sentiment)\n",
    "    predicted_label = (torch.sigmoid(outputs.logits)>0.5).float()\n",
    "    # Return predicted sentiment label\n",
    "    return 'positive' if predicted_label == 1 else 'negative'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "positive\n"
     ]
    }
   ],
   "source": [
    "# Example usage\n",
    "text = \"I love this product! It's amazing!\"\n",
    "sentiment = get_sentiment(text)\n",
    "print(sentiment) # Output: 'positive'"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "base",
   "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": "3.11.4"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}