{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "2d4667c5",
   "metadata": {},
   "outputs": [],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "524483a8",
   "metadata": {},
   "source": [
    "# Dependencies"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "4c51929f",
   "metadata": {},
   "outputs": [],
   "source": [
    "import sys\n",
    "sys.path.append('..')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "8f19f1c8",
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# torch\n",
    "import torch\n",
    "import torchaudio\n",
    "from torch.utils.data import DataLoader\n",
    "\n",
    "# model training\n",
    "from cnn import CNNetwork\n",
    "from dataset import VoiceDataset\n",
    "from train import train"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4f51b4f8",
   "metadata": {},
   "source": [
    "## Globals"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "6f1716bf",
   "metadata": {},
   "outputs": [],
   "source": [
    "DATA_PATH = os.path.join('..', 'data')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "13fa6700",
   "metadata": {},
   "outputs": [],
   "source": [
    "TRAIN_PATH = os.path.join(DATA_PATH, 'train')\n",
    "TEST_PATH = os.path.join(DATA_PATH, 'test')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "152c1fbf",
   "metadata": {},
   "outputs": [],
   "source": [
    "EPOCHS = 100\n",
    "BATCH_SIZE = 128\n",
    "LEARNING_RATE = 0.001\n",
    "SAMPLE_RATE=16000"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "98cf668e",
   "metadata": {},
   "outputs": [],
   "source": [
    "MEL_SPEC = torchaudio.transforms.MelSpectrogram(\n",
    "    n_fft=1024,\n",
    "    hop_length=512,\n",
    "    n_mels=64,\n",
    "    sample_rate=SAMPLE_RATE,\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "063c13a4",
   "metadata": {},
   "source": [
    "# `train()` setup"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "19b58a13",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using cpu device.\n"
     ]
    }
   ],
   "source": [
    "if torch.cuda.is_available():\n",
    "    device = \"cuda\"\n",
    "else:\n",
    "    device = \"cpu\"\n",
    "print(f\"Using {device} device.\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "848745e7",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Datasets\n",
    "train_dataset = VoiceDataset(TRAIN_PATH, MEL_SPEC, device, SAMPLE_RATE)\n",
    "test_dataset = VoiceDataset(TEST_PATH, MEL_SPEC, device, SAMPLE_RATE)\n",
    "\n",
    "# Dataloaders\n",
    "train_dataloader = DataLoader(train_dataset, batch_size=BATCH_SIZE, shuffle=True)\n",
    "test_dataloader = DataLoader(test_dataset, batch_size=BATCH_SIZE, shuffle=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "67a76338",
   "metadata": {},
   "outputs": [],
   "source": [
    "model = CNNetwork().to(device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "5c5b08c3",
   "metadata": {},
   "outputs": [],
   "source": [
    "loss_fn = torch.nn.CrossEntropyLoss()\n",
    "optimizer = torch.optim.Adam(model.parameters(), lr=LEARNING_RATE)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e6eb6722",
   "metadata": {},
   "source": [
    "# Train/Test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "f4b5c789",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:   0%|                                                           | 0/100 [00:00<?, ?it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:   1%|▍                                                | 1/100 [00:45<1:15:43, 45.89s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.73, Training Accuracy  0.84\n",
      "Testing Loss: 0.71, Testing Accuracy  0.84\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 2\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:   2%|▉                                                | 2/100 [01:31<1:14:30, 45.61s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.73, Training Accuracy  0.82\n",
      "Testing Loss: 0.76, Testing Accuracy  0.79\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 3\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:   3%|█▍                                               | 3/100 [02:16<1:13:38, 45.55s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.77, Training Accuracy  0.78\n",
      "Testing Loss: 0.71, Testing Accuracy  0.84\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 4\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:   4%|█▉                                               | 4/100 [03:02<1:12:43, 45.46s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.73, Training Accuracy  0.82\n",
      "Testing Loss: 0.66, Testing Accuracy  0.89\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 5\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:   5%|██▍                                              | 5/100 [03:47<1:11:51, 45.38s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.65, Training Accuracy  0.90\n",
      "Testing Loss: 0.71, Testing Accuracy  0.84\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 6\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:   6%|██▉                                              | 6/100 [04:32<1:11:05, 45.37s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.65, Training Accuracy  0.90\n",
      "Testing Loss: 0.55, Testing Accuracy  1.00\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 7\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:   7%|███▍                                             | 7/100 [05:17<1:10:09, 45.27s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.60, Training Accuracy  0.95\n",
      "Testing Loss: 0.55, Testing Accuracy  1.00\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 8\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:   8%|███▉                                             | 8/100 [06:03<1:09:29, 45.32s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.59, Training Accuracy  0.96\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 9\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:   9%|████▍                                            | 9/100 [06:48<1:08:47, 45.36s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.59, Training Accuracy  0.96\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 10\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  10%|████▊                                           | 10/100 [07:34<1:08:08, 45.42s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.59, Training Accuracy  0.96\n",
      "Testing Loss: 0.55, Testing Accuracy  1.00\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 11\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  11%|█████▎                                          | 11/100 [08:19<1:07:23, 45.43s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.58, Training Accuracy  0.97\n",
      "Testing Loss: 0.55, Testing Accuracy  1.00\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 12\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  12%|█████▊                                          | 12/100 [09:05<1:07:02, 45.71s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.58, Training Accuracy  0.97\n",
      "Testing Loss: 0.71, Testing Accuracy  0.84\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 13\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  13%|██████▏                                         | 13/100 [09:52<1:06:43, 46.02s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.61, Training Accuracy  0.94\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 14\n",
      "Training Loss: 0.59, Training Accuracy  0.96\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  14%|██████▋                                         | 14/100 [10:52<1:12:04, 50.28s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Testing Loss: 0.55, Testing Accuracy  1.00\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 15\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  15%|███████                                        | 15/100 [16:47<3:21:11, 142.02s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.58, Training Accuracy  0.97\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 16\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  16%|███████▌                                       | 16/100 [17:38<2:40:16, 114.49s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.58, Training Accuracy  0.97\n",
      "Testing Loss: 0.55, Testing Accuracy  1.00\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 17\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  17%|████████▏                                       | 17/100 [18:25<2:10:31, 94.35s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.58, Training Accuracy  0.97\n",
      "Testing Loss: 0.55, Testing Accuracy  1.00\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 18\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  18%|████████▋                                       | 18/100 [19:13<1:49:46, 80.32s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.58, Training Accuracy  0.98\n",
      "Testing Loss: 0.55, Testing Accuracy  1.00\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 19\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  19%|█████████                                       | 19/100 [20:00<1:35:02, 70.40s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.58, Training Accuracy  0.97\n",
      "Testing Loss: 0.55, Testing Accuracy  1.00\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 20\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  20%|█████████▌                                      | 20/100 [20:47<1:24:38, 63.48s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.59, Training Accuracy  0.96\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 21\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  21%|██████████                                      | 21/100 [21:35<1:17:10, 58.62s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.64, Training Accuracy  0.91\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 22\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  22%|██████████▌                                     | 22/100 [22:23<1:12:04, 55.44s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.59, Training Accuracy  0.96\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 23\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  23%|███████████                                     | 23/100 [23:11<1:08:20, 53.25s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.57, Training Accuracy  0.98\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 24\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  24%|███████████▌                                    | 24/100 [23:57<1:04:54, 51.25s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.57, Training Accuracy  0.98\n",
      "Testing Loss: 0.55, Testing Accuracy  1.00\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 25\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  25%|████████████                                    | 25/100 [24:44<1:02:18, 49.84s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.57, Training Accuracy  0.98\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 26\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  26%|████████████▍                                   | 26/100 [25:31<1:00:17, 48.88s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.57, Training Accuracy  0.99\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 27\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  27%|█████████████▌                                    | 27/100 [26:17<58:38, 48.20s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  0.99\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 28\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  28%|██████████████                                    | 28/100 [27:04<57:20, 47.78s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  0.99\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 29\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  29%|█████████████▉                                  | 29/100 [28:51<1:17:25, 65.43s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  0.99\n",
      "Testing Loss: 0.55, Testing Accuracy  1.00\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 30\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  30%|██████████████                                 | 30/100 [42:13<5:34:16, 286.52s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  0.99\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 31\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  31%|██████████████▌                                | 31/100 [42:59<4:06:37, 214.46s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  0.99\n",
      "Testing Loss: 0.55, Testing Accuracy  1.00\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 32\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  32%|███████████████                                | 32/100 [43:45<3:05:47, 163.94s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  0.99\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 33\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  33%|███████████████▌                               | 33/100 [44:32<2:23:41, 128.67s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.57, Training Accuracy  0.98\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 34\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  34%|███████████████▉                               | 34/100 [45:18<1:54:26, 104.04s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.57, Training Accuracy  0.98\n",
      "Testing Loss: 0.55, Testing Accuracy  1.00\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 35\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  35%|████████████████▊                               | 35/100 [46:05<1:33:56, 86.71s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.57, Training Accuracy  0.98\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 36\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  36%|█████████████████▎                              | 36/100 [46:51<1:19:35, 74.62s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.62, Training Accuracy  0.93\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 37\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  37%|█████████████████▊                              | 37/100 [47:37<1:09:28, 66.17s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.59, Training Accuracy  0.96\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 38\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  38%|██████████████████▏                             | 38/100 [48:24<1:02:15, 60.24s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.57, Training Accuracy  0.98\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 39\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  39%|███████████████████▌                              | 39/100 [49:11<57:10, 56.24s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  0.99\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 40\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  40%|████████████████████                              | 40/100 [49:57<53:20, 53.35s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.57, Training Accuracy  0.99\n",
      "Testing Loss: 0.55, Testing Accuracy  1.00\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 41\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  41%|████████████████████▌                             | 41/100 [50:43<50:19, 51.18s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.58, Training Accuracy  0.97\n",
      "Testing Loss: 0.55, Testing Accuracy  1.00\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 42\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  42%|█████████████████████                             | 42/100 [51:30<48:04, 49.73s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  0.99\n",
      "Testing Loss: 0.55, Testing Accuracy  1.00\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 43\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  43%|█████████████████████▌                            | 43/100 [52:16<46:10, 48.61s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  0.99\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 44\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  44%|████████████████████▋                          | 44/100 [58:22<2:14:12, 143.79s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  0.99\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 45\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  45%|████████████████████▎                        | 45/100 [1:14:56<6:05:37, 398.86s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  0.99\n",
      "Testing Loss: 0.55, Testing Accuracy  1.00\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 46\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  46%|████████████████████▋                        | 46/100 [1:15:48<4:25:15, 294.73s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  0.99\n",
      "Testing Loss: 0.55, Testing Accuracy  1.00\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 47\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  47%|█████████████████████▏                       | 47/100 [1:16:37<3:15:15, 221.05s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  0.99\n",
      "Testing Loss: 0.55, Testing Accuracy  1.00\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 48\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  48%|█████████████████████▌                       | 48/100 [1:17:27<2:27:04, 169.71s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  0.99\n",
      "Testing Loss: 0.55, Testing Accuracy  1.00\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 49\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  49%|██████████████████████                       | 49/100 [1:18:15<1:53:23, 133.41s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  0.99\n",
      "Testing Loss: 0.55, Testing Accuracy  1.00\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 50\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  50%|██████████████████████▌                      | 50/100 [1:19:04<1:30:06, 108.12s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  0.99\n",
      "Testing Loss: 0.55, Testing Accuracy  1.00\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 51\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  51%|███████████████████████▍                      | 51/100 [1:19:53<1:13:39, 90.18s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  0.99\n",
      "Testing Loss: 0.55, Testing Accuracy  1.00\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 52\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  52%|███████████████████████▉                      | 52/100 [1:20:43<1:02:36, 78.25s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  0.99\n",
      "Testing Loss: 0.55, Testing Accuracy  1.00\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 53\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  53%|█████████████████████████▍                      | 53/100 [1:21:34<54:50, 70.00s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  0.99\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 54\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  54%|█████████████████████████▉                      | 54/100 [1:22:23<48:58, 63.89s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  0.99\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 55\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  55%|██████████████████████████▍                     | 55/100 [1:23:13<44:39, 59.54s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 56\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  56%|██████████████████████████▉                     | 56/100 [1:24:04<41:47, 57.00s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 57\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  57%|███████████████████████████▎                    | 57/100 [1:24:54<39:16, 54.80s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 58\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  58%|███████████████████████████▊                    | 58/100 [1:25:45<37:42, 53.87s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 59\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  59%|████████████████████████████▎                   | 59/100 [1:26:33<35:35, 52.09s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 60\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  60%|████████████████████████████▊                   | 60/100 [1:27:20<33:44, 50.61s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 61\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  61%|█████████████████████████████▎                  | 61/100 [1:28:07<32:06, 49.40s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 62\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  62%|█████████████████████████████▊                  | 62/100 [1:28:55<31:04, 49.06s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 63\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  63%|████████████████████████████▎                | 63/100 [1:34:40<1:24:59, 137.81s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 64\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  64%|████████████████████████████▊                | 64/100 [1:36:13<1:14:31, 124.21s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 65\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  65%|██████████████████████████████▌                | 65/100 [1:36:59<58:45, 100.72s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 66\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  66%|███████████████████████████████▋                | 66/100 [1:37:44<47:44, 84.24s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 67\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  67%|████████████████████████████████▏               | 67/100 [1:38:30<39:57, 72.66s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 68\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  68%|████████████████████████████████▋               | 68/100 [1:39:15<34:23, 64.50s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 69\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  69%|█████████████████████████████████               | 69/100 [1:40:01<30:24, 58.84s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 70\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  70%|█████████████████████████████████▌              | 70/100 [1:40:47<27:28, 54.94s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 71\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  71%|██████████████████████████████████              | 71/100 [1:41:33<25:16, 52.30s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 72\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  72%|██████████████████████████████████▌             | 72/100 [1:42:19<23:32, 50.44s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 73\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  73%|███████████████████████████████████             | 73/100 [1:43:05<22:06, 49.12s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 74\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  74%|███████████████████████████████████▌            | 74/100 [1:43:51<20:54, 48.24s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 75\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  75%|████████████████████████████████████            | 75/100 [1:44:38<19:53, 47.73s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 76\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  76%|████████████████████████████████████▍           | 76/100 [1:45:25<18:58, 47.45s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 77\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  77%|████████████████████████████████████▉           | 77/100 [1:46:15<18:32, 48.38s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 78\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  78%|███████████████████████████████████          | 78/100 [2:03:08<2:03:52, 337.83s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 79\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  79%|███████████████████████████████████▌         | 79/100 [2:21:11<3:16:23, 561.12s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 80\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  80%|████████████████████████████████████         | 80/100 [2:26:56<2:45:26, 496.31s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 81\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  81%|████████████████████████████████████▍        | 81/100 [2:32:40<2:22:42, 450.67s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 82\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  82%|████████████████████████████████████▉        | 82/100 [2:33:26<1:38:46, 329.24s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 83\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  83%|█████████████████████████████████████▎       | 83/100 [2:39:10<1:34:32, 333.69s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 84\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  84%|█████████████████████████████████████▊       | 84/100 [2:55:21<2:19:58, 524.89s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 85\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  85%|██████████████████████████████████████▎      | 85/100 [3:11:28<2:44:22, 657.47s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 86\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  86%|██████████████████████████████████████▋      | 86/100 [3:29:05<3:01:23, 777.36s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 87\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  87%|███████████████████████████████████████▏     | 87/100 [3:45:40<3:02:34, 842.66s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 88\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  88%|███████████████████████████████████████▌     | 88/100 [3:46:25<2:00:42, 603.57s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 89\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  89%|████████████████████████████████████████     | 89/100 [3:47:11<1:19:58, 436.20s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 90\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  90%|██████████████████████████████████████████▎    | 90/100 [3:47:57<53:10, 319.00s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 91\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  91%|██████████████████████████████████████████▊    | 91/100 [3:48:43<35:35, 237.23s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 92\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  92%|███████████████████████████████████████████▏   | 92/100 [3:49:29<23:59, 179.94s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 93\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  93%|███████████████████████████████████████████▋   | 93/100 [3:50:16<16:19, 139.92s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 94\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  94%|████████████████████████████████████████████▏  | 94/100 [3:51:02<11:11, 111.90s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 95\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  95%|█████████████████████████████████████████████▌  | 95/100 [3:51:49<07:41, 92.29s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 96\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  96%|██████████████████████████████████████████████  | 96/100 [3:52:35<05:14, 78.51s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 97\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  97%|██████████████████████████████████████████████▌ | 97/100 [3:53:22<03:26, 68.89s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 98\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  98%|███████████████████████████████████████████████ | 98/100 [3:54:09<02:04, 62.29s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 99\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\r\n",
      "Training model...:  99%|███████████████████████████████████████████████▌| 99/100 [3:54:55<00:57, 57.51s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "Epoch 100\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training model...: 100%|██████████████████████████████████████████████| 100/100 [3:55:42<00:00, 141.42s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training Loss: 0.56, Training Accuracy  1.00\n",
      "Testing Loss: 0.60, Testing Accuracy  0.95\n",
      "-------------------------------------------- \n",
      "\n",
      "---- Finished Training ----\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "history = train(\n",
    "    model,\n",
    "    train_dataloader,\n",
    "    loss_fn,\n",
    "    optimizer,\n",
    "    device,\n",
    "    EPOCHS,\n",
    "    test_dataloader\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "572ae86d",
   "metadata": {},
   "source": [
    "# Visualizations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "0b986ac4",
   "metadata": {},
   "outputs": [],
   "source": [
    "training_acc = history[0]\n",
    "training_loss = history[1]\n",
    "\n",
    "testing_acc = history[2]\n",
    "testing_loss = history[3]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "5cfbedcc",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure()\n",
    "fig.suptitle(\"Accuracy\")\n",
    "ax = fig.subplots(1, 2)\n",
    "ax[0].plot(training_acc)\n",
    "ax[0].title.set_text(\"Training\")\n",
    "ax[1].plot(testing_acc)\n",
    "ax[1].title.set_text(\"Testing\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "2475aebc",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure()\n",
    "fig.suptitle(\"Loss\")\n",
    "ax = fig.subplots(1, 2)\n",
    "ax[0].plot(training_loss)\n",
    "ax[0].title.set_text(\"Training\")\n",
    "ax[1].plot(testing_loss)\n",
    "ax[1].title.set_text(\"Testing\")"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.16"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}