train.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "44ddf287-4125-4d02-a3db-09a4f358b8f3",
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import seaborn as sns\n",
    "\n",
    "import random\n",
    "from glob import glob\n",
    "from tqdm import tqdm\n",
    "from scipy.io import loadmat\n",
    "\n",
    "import torch\n",
    "from torch import nn\n",
    "from torch.optim import Adam\n",
    "from torch.optim.lr_scheduler import CosineAnnealingLR\n",
    "from torch.utils.data import Dataset, DataLoader\n",
    "import torch.nn.functional as F"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "e2868d4b-ebc6-40e5-8242-80fdb12670df",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['TrainingSet3', 'Label.csv', 'alldata', 'TrainingSet1', 'single_label.csv', 'TrainingSet2']\n"
     ]
    }
   ],
   "source": [
    "data_dir = \"/media/mountHDD2/khoibaocon\"\n",
    "print(os.listdir(data_dir))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "2f3c37b5-d09e-467c-8ee3-c2bb3fce5555",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(6877, 4)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "main_df = pd.read_csv(data_dir + \"/Label.csv\")\n",
    "main_df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "a4f1e0f7-99a0-4a5a-8ba5-7cab7e5d354b",
   "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>Recording</th>\n",
       "      <th>First_label</th>\n",
       "      <th>Second_label</th>\n",
       "      <th>Third_label</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>A0001</td>\n",
       "      <td>5</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>A0002</td>\n",
       "      <td>1</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>A0003</td>\n",
       "      <td>2</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>A0004</td>\n",
       "      <td>2</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>A0005</td>\n",
       "      <td>7</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  Recording  First_label  Second_label  Third_label\n",
       "0     A0001            5           NaN          NaN\n",
       "1     A0002            1           NaN          NaN\n",
       "2     A0003            2           NaN          NaN\n",
       "3     A0004            2           NaN          NaN\n",
       "4     A0005            7           NaN          NaN"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "main_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "e460e120-cf80-48b0-b2d6-cfbfc30cc08c",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "First_label\n",
       "5    1695\n",
       "2    1098\n",
       "1     918\n",
       "8     826\n",
       "3     704\n",
       "7     653\n",
       "6     574\n",
       "4     207\n",
       "9     202\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "main_df[\"First_label\"].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "946947c7-a713-4ba1-8184-2fef44eba6cc",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Second_label\n",
       "5.0    162\n",
       "2.0    123\n",
       "7.0     47\n",
       "6.0     42\n",
       "8.0     41\n",
       "4.0     28\n",
       "3.0     18\n",
       "9.0     16\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "main_df[\"Second_label\"].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "982de563-1cc2-44d7-8ff8-f8096b9e7115",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Third_label\n",
       "9.0    2\n",
       "8.0    2\n",
       "6.0    1\n",
       "4.0    1\n",
       "Name: count, dtype: int64"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "main_df[\"Third_label\"].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "e55bab59-4461-434b-9f39-ca1413ae7ed6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(6400, 4)"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "single_main_df = main_df[main_df[\"Second_label\"].isnull()]\n",
    "single_main_df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "7c439e8c-8f9c-4237-9e6a-c0939efe403b",
   "metadata": {},
   "outputs": [],
   "source": [
    "single_main_df.to_csv(data_dir + \"/single_label.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "069eddeb-73ab-40de-8a4f-2b4cae764db6",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "6877\n"
     ]
    }
   ],
   "source": [
    "mat_files = glob(data_dir + \"/alldata/*\")\n",
    "print(len(mat_files))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "6235c55f-7003-4eba-9dc7-a7c107721c22",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "6400\n"
     ]
    }
   ],
   "source": [
    "single_fns = single_main_df[\"Recording\"].values.tolist()\n",
    "print(len(single_fns))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "cf74da3f-b040-4514-b9e3-642b8c501dba",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "True\n"
     ]
    }
   ],
   "source": [
    "single_mat_paths = [data_dir + f\"/alldata/{x}.mat\" for x in single_fns]\n",
    "print(os.path.exists(single_mat_paths[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "bb59505a-ddb3-4835-8673-354e15b764e9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dict_keys(['__header__', '__version__', '__globals__', 'ECG'])"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sample_data = loadmat(single_mat_paths[0])\n",
    "sample_data.keys()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "db6cff39-d48b-49d3-80c7-e9d6888b62dd",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(12, 7500)"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "sample_signal_data = sample_data['ECG'][0][0][2]\n",
    "sample_signal_data.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "c74162e9-da4b-43fe-802c-59e884bf9f98",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7fd58be30490>]"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(sample_signal_data[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "7978703e-8bd0-464d-a4d5-08abc919c38a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([1, 12, 32])\n"
     ]
    }
   ],
   "source": [
    "sample_sig = torch.randn(1, 12, 32)\n",
    "conv_test = nn.Conv1d(12, 12, 3, 1, 1)\n",
    "print(conv_test(sample_sig).shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "id": "b553864b-76e1-4a92-aa92-ecb7410840ac",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "MAX: 72000\n",
      "MIN: 3000\n",
      "AVG: 7946.03703125\n"
     ]
    }
   ],
   "source": [
    "len_lst = [loadmat(x)['ECG'][0][0][2].shape[1] for x in single_mat_paths]\n",
    "print(f\"MAX: {max(len_lst)}\")\n",
    "print(f\"MIN: {min(len_lst)}\")\n",
    "print(f\"AVG: {sum(len_lst)/len(len_lst)}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "id": "949a0ef3-5be8-4e8e-b087-fa05e95f6fb8",
   "metadata": {},
   "outputs": [],
   "source": [
    "class BasicBlock(nn.Module):BasicBlock\n",
    "    def __init__(self, channel_num):\n",
    "        super(BasicBlock, self).__init__()\n",
    "        self.conv_block1 = nn.Sequential(\n",
    "\t\t\tnn.Conv1d(channel_num, channel_num, 3, padding=1),\n",
    "\t\t\tnn.BatchNorm1d(channel_num),\n",
    "\t\t\tnn.ReLU(),\n",
    "\t\t)\n",
    "        self.conv_block2 = nn.Sequential(\n",
    "\t\t\tnn.Conv1d(channel_num, channel_num, 3, padding=1),\n",
    "\t\t\tnn.BatchNorm1d(channel_num),\n",
    "\t\t)\n",
    "        self.relu = nn.ReLU()\n",
    "        torch.nn.init.kaiming_normal_(self.conv_block1[0].weight)\n",
    "        torch.nn.init.kaiming_normal_(self.conv_block2[0].weight)\n",
    "        \n",
    "    def forward(self, x):\n",
    "        residual = x\n",
    "        x = self.conv_block1(x)\n",
    "        x = self.conv_block2(x)\n",
    "        x = x + residual\n",
    "        out = self.relu(x)\n",
    "        return out"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "id": "dbdfa702-b20a-4fe3-80e6-e28fa4ac0e6f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([1, 2, 32])\n"
     ]
    }
   ],
   "source": [
    "test_basic_block = BasicBlock(2)\n",
    "sample_sig = torch.randn(1, 2, 32)\n",
    "print(test_basic_block(sample_sig).shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "id": "e07a7549-965d-49b4-b498-ff00f4014167",
   "metadata": {},
   "outputs": [],
   "source": [
    "class ResNet(nn.Module):\n",
    "    def __init__(self, in_channels = 12, type = 18, num_classes = 9):\n",
    "        super(ResNet, self).__init__()\n",
    "        self.struc_dict = {\n",
    "            18: {\n",
    "                \"num_channels\" : [64, 128, 256, 512],\n",
    "                \"counts\" : [2, 2, 2, 2]\n",
    "            }\n",
    "        }\n",
    "        self.conv1 = nn.Conv1d(in_channels=in_channels, out_channels=64, kernel_size=7, stride=2)\n",
    "        torch.nn.init.kaiming_normal_(self.conv1.weight)\n",
    "        self.max1 = nn.MaxPool1d(kernel_size=3, stride=2)\n",
    "        self.main = nn.Sequential()\n",
    "        for idx, struc in enumerate(\n",
    "            zip(\n",
    "                self.struc_dict[type][\"num_channels\"], \n",
    "                self.struc_dict[type][\"counts\"]\n",
    "            )\n",
    "        ):\n",
    "            num_channel, cnt = struc\n",
    "            for i in range(cnt):\n",
    "                self.main.add_module(f\"conv{idx+1}_{i}\", BasicBlock(num_channel))\n",
    "            if idx < len(self.struc_dict[type][\"num_channels\"]) - 1:\n",
    "                self.main.add_module(f\"ext_{idx}\", nn.Conv1d(num_channel, self.struc_dict[type][\"num_channels\"][idx+1], 3, 1))\n",
    "                self.main.add_module(f\"extbn_{idx}\", nn.BatchNorm1d(self.struc_dict[type][\"num_channels\"][idx+1]))\n",
    "                                     \n",
    "        self.avg = torch.nn.AdaptiveAvgPool1d((1))\n",
    "        self.lin = nn.Linear(self.struc_dict[type][\"num_channels\"][-1], num_classes)\n",
    "        torch.nn.init.kaiming_normal_(self.lin.weight)\n",
    "    def forward(self, x):\n",
    "        x = self.conv1(x)\n",
    "        x = self.max1(x)\n",
    "        x = self.main(x)\n",
    "        x = self.avg(x)\n",
    "        x = x.reshape(x.shape[0], -1)\n",
    "        x = self.lin(x)\n",
    "        return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "id": "182b31c1-94b6-4c83-837e-f8aa593a1324",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([1, 9])"
      ]
     },
     "execution_count": 123,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model = ResNet()\n",
    "sample_sig = torch.randn(1, 12, 3000)\n",
    "model(sample_sig).shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 143,
   "id": "c7dd7940-7c95-4edf-b39e-0de905f24dc5",
   "metadata": {},
   "outputs": [],
   "source": [
    "class ECG(Dataset):\n",
    "    def __init__(self, data_paths, label_df):\n",
    "        self.data_paths = data_paths\n",
    "        random.shuffle(self.data_paths)\n",
    "        self.label_df = label_df\n",
    "\n",
    "    def __getitem__(self, idx):\n",
    "        data_path = self.data_paths[idx]        \n",
    "        data = loadmat(data_path)['ECG'][0][0][2]\n",
    "        clip_data = data[:, 300:3000]\n",
    "\n",
    "        filename = data_path.split(\"/\")[-1].split(\".\")[0]\n",
    "        label = self.label_df[self.label_df[\"Recording\"] == filename][\"First_label\"].values.item()\n",
    "\n",
    "        torch_data = torch.from_numpy(clip_data)\n",
    "\n",
    "        return torch_data.float(), label-1\n",
    "\n",
    "    def __len__(self):\n",
    "        return len(self.data_paths)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 144,
   "id": "4e01d387-99b1-416e-83a7-cb1d3a664179",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([12, 2700]) 0\n"
     ]
    }
   ],
   "source": [
    "check_ds = ECG(data_paths=single_mat_paths, label_df=single_main_df)\n",
    "sample, lbl = check_ds[0]\n",
    "print(sample.shape, lbl)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "id": "68e75806-1788-4302-a883-2a1e6401ea00",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([1, 9])"
      ]
     },
     "execution_count": 145,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model(sample.unsqueeze(dim=0)).shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "id": "07cf3c4b-a69a-4c74-8d60-a841345c04a0",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0->918\n",
      "1->976\n",
      "2->686\n",
      "3->179\n",
      "4->1533\n",
      "5->532\n",
      "6->607\n",
      "7->784\n",
      "8->185\n"
     ]
    }
   ],
   "source": [
    "data_dict = {\n",
    "    idx : [] for idx in range(9)\n",
    "}\n",
    "\n",
    "for data_path in single_mat_paths:\n",
    "    filename = data_path.split(\"/\")[-1].split(\".\")[0]\n",
    "    _cls = single_main_df[single_main_df[\"Recording\"] == filename][\"First_label\"].values.item()\n",
    "\n",
    "    data_dict[_cls-1].append(data_path)\n",
    "\n",
    "for key in data_dict:\n",
    "    print(f\"{key}->{len(data_dict[key])}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 148,
   "id": "ef039dff-5ef3-4f80-9d7b-df982fe4b0f5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0->826--92\n",
      "1->878--98\n",
      "2->617--69\n",
      "3->161--18\n",
      "4->1379--154\n",
      "5->478--54\n",
      "6->546--61\n",
      "7->705--79\n",
      "8->166--19\n"
     ]
    }
   ],
   "source": [
    "train_data_dict = {\n",
    "    _cls : data_dict[_cls][:int(0.9*len(data_dict[_cls]))] for _cls in data_dict\n",
    "}\n",
    "\n",
    "valid_data_dict = {\n",
    "    _cls : data_dict[_cls][int(0.9*len(data_dict[_cls])):] for _cls in data_dict\n",
    "}\n",
    "\n",
    "for key in train_data_dict:\n",
    "    print(f\"{key}->{len(train_data_dict[key])}--{len(valid_data_dict[key])}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 154,
   "id": "4fa08ec9-2dc6-4690-ab1b-f653d3a8d10d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "5756\n",
      "644\n"
     ]
    }
   ],
   "source": [
    "train_data_paths = []\n",
    "for key in train_data_dict:\n",
    "    train_data_paths.extend(train_data_dict[key])\n",
    "valid_data_paths = []\n",
    "for key in valid_data_dict:\n",
    "    valid_data_paths.extend(valid_data_dict[key])\n",
    "print(len(train_data_paths))\n",
    "print(len(valid_data_paths))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "id": "5e92b032-fb47-4b7c-b675-9ea3ae5d8a45",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "5756\n",
      "644\n"
     ]
    }
   ],
   "source": [
    "train_ds = ECG(train_data_paths, single_main_df)\n",
    "valid_ds = ECG(valid_data_paths, single_main_df)\n",
    "\n",
    "print(len(train_ds))\n",
    "print(len(valid_ds))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 157,
   "id": "701f345f-bc5e-434d-ab2f-ece864f14b99",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "180\n",
      "644\n"
     ]
    }
   ],
   "source": [
    "device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\", index = 1)\n",
    "batch_size = 32\n",
    "\n",
    "traindl = DataLoader(\n",
    "    train_ds,\n",
    "    batch_size=batch_size, \n",
    "    shuffle=True, \n",
    "    pin_memory=True, \n",
    "    num_workers=os.cpu_count()//2\n",
    ")\n",
    "\n",
    "validdl = DataLoader(\n",
    "    valid_ds,\n",
    "    batch_size=1, \n",
    "    shuffle=True, \n",
    "    pin_memory=True, \n",
    "    num_workers=os.cpu_count()//2\n",
    ")\n",
    "\n",
    "print(len(traindl))\n",
    "print(len(validdl))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 159,
   "id": "0e3fd1ae-3394-4ecb-9681-c34210dcbd05",
   "metadata": {},
   "outputs": [],
   "source": [
    "epoch = 100\n",
    "lr = 0.0001\n",
    "\n",
    "model.to(device)\n",
    "optimizer = Adam(model.parameters(), lr=lr)\n",
    "scheduler = CosineAnnealingLR(optimizer=optimizer, T_max=epoch)\n",
    "loss_fn = nn.CrossEntropyLoss()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 161,
   "id": "d172aded-4332-40ad-9206-d2ea45c5b545",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch: 0\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "180it [00:08, 21.54it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 1.2191842674543072 - train acc: 60.215427380125085\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 342.01it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.193489576425724 - valid acc: 61.49068322981367\n",
      "Epoch: 1\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.60it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 1.1113728774326475 - train acc: 63.429464906184855\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 340.01it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.0930871936171211 - valid acc: 64.44099378881988\n",
      "Epoch: 2\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.52it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 1.0170088883218817 - train acc: 66.73036831132731\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 346.58it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.0159193276765957 - valid acc: 66.77018633540372\n",
      "Epoch: 3\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.41it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.9080482456604195 - train acc: 70.34398888116748\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 323.64it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 0.9131319369071985 - valid acc: 69.72049689440993\n",
      "Epoch: 4\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.43it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.8108531366513428 - train acc: 73.59277275886032\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 327.73it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 0.8960865774456034 - valid acc: 68.94409937888199\n",
      "Epoch: 5\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.44it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.7572465277583905 - train acc: 75.05211952744962\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 337.39it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 0.8529946182070673 - valid acc: 72.04968944099379\n",
      "Epoch: 6\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.39it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.6991573578818551 - train acc: 77.72758860319666\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 334.45it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 0.89890343362997 - valid acc: 71.42857142857143\n",
      "Epoch: 7\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "180it [00:08, 21.45it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.6594170480134101 - train acc: 78.28353022932592\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 355.19it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 0.9581760890762877 - valid acc: 70.4968944099379\n",
      "Epoch: 8\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.39it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.6480723890845336 - train acc: 79.04794996525365\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 326.47it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 0.8897916690314391 - valid acc: 70.1863354037267\n",
      "Epoch: 9\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.43it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.61912993337522 - train acc: 80.31619179986102\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 345.00it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 0.8901127665841254 - valid acc: 71.73913043478261\n",
      "Epoch: 10\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.38it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.6036179305454872 - train acc: 80.68102849200834\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 340.46it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 0.834571241282156 - valid acc: 73.13664596273291\n",
      "Epoch: 11\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.28it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.6014086680705321 - train acc: 79.62126476719943\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 344.44it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 0.789757848052782 - valid acc: 74.68944099378882\n",
      "Epoch: 12\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.42it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.5676643717721854 - train acc: 81.61917998610147\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 345.07it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 0.7704657316518331 - valid acc: 75.0\n",
      "Epoch: 13\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.42it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.5513888642441627 - train acc: 82.17512161223071\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 336.91it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 0.7976105575342348 - valid acc: 74.84472049689441\n",
      "Epoch: 14\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.35it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.5173583197194105 - train acc: 82.88742182070882\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 344.93it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 0.7890755982840023 - valid acc: 74.53416149068323\n",
      "Epoch: 15\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.35it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.4896886247995845 - train acc: 84.36414176511467\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 342.52it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 0.8614605929613541 - valid acc: 73.4472049689441\n",
      "Epoch: 16\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.42it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.4724771291350519 - train acc: 84.71160528144544\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 339.34it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 0.8982661005908604 - valid acc: 72.67080745341616\n",
      "Epoch: 17\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.43it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.42958368760580457 - train acc: 85.94510076441973\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 341.94it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 0.9967525795467017 - valid acc: 71.73913043478261\n",
      "Epoch: 18\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.41it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.413516090985117 - train acc: 86.58790826963168\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 336.62it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.3842643792813307 - valid acc: 56.2111801242236\n",
      "Epoch: 19\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.37it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.4226514176782949 - train acc: 86.44892286309937\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 337.17it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 0.9981345504479263 - valid acc: 73.13664596273291\n",
      "Epoch: 20\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.32it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.39250740758533587 - train acc: 87.14384989576095\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 345.82it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 0.8363274216142099 - valid acc: 72.20496894409938\n",
      "Epoch: 21\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.39it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.38656964377817493 - train acc: 87.57817929117444\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 339.10it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 0.8424210242633916 - valid acc: 74.68944099378882\n",
      "Epoch: 22\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.36it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.3894054406015567 - train acc: 87.19596942321056\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 337.69it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 0.8348378496556836 - valid acc: 75.46583850931677\n",
      "Epoch: 23\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.38it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.362053664827813 - train acc: 88.20361362056984\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 338.76it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 0.8060450176062167 - valid acc: 75.0\n",
      "Epoch: 24\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.22it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.3289315264526359 - train acc: 89.10701876302987\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 343.60it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 0.8310469828724465 - valid acc: 73.75776397515527\n",
      "Epoch: 25\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.33it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.3201579762487438 - train acc: 89.61084086170952\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 325.49it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 0.8922432038026634 - valid acc: 73.91304347826086\n",
      "Epoch: 26\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.25it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.29816757325686555 - train acc: 90.09728978457262\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 338.80it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 0.9941214162846703 - valid acc: 72.36024844720497\n",
      "Epoch: 27\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.30it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.278530868238577 - train acc: 91.00069492703267\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 336.64it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.5741614079979482 - valid acc: 63.975155279503106\n",
      "Epoch: 28\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.27it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.2587924328263245 - train acc: 91.93884642112579\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 343.27it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.2179523947370454 - valid acc: 67.70186335403726\n",
      "Epoch: 29\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.26it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.26125905887541157 - train acc: 92.02571230020847\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 338.27it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.0356445621845345 - valid acc: 70.4968944099379\n",
      "Epoch: 30\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.32it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.2637338577951799 - train acc: 91.53926337734538\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 341.30it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 0.9502230782061907 - valid acc: 72.51552795031056\n",
      "Epoch: 31\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.27it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.2553098510805122 - train acc: 91.73036831132731\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 350.11it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 0.8719901056918554 - valid acc: 74.68944099378882\n",
      "Epoch: 32\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.29it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.22692058002316087 - train acc: 92.72063933287005\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 333.86it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 0.8700304283242654 - valid acc: 73.29192546583852\n",
      "Epoch: 33\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.30it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.22793354437421154 - train acc: 92.68589298123698\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 336.11it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 0.9000923583808504 - valid acc: 73.29192546583852\n",
      "Epoch: 34\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.36it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.23691129301513372 - train acc: 92.26893676164003\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 333.99it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 0.8988638788469896 - valid acc: 74.37888198757764\n",
      "Epoch: 35\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.29it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.20697073655088521 - train acc: 93.363446838082\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 333.88it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.0407632212759406 - valid acc: 72.20496894409938\n",
      "Epoch: 36\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "180it [00:08, 21.28it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.1741263039680666 - train acc: 94.7359277275886\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 347.72it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.195857472851632 - valid acc: 70.34161490683229\n",
      "Epoch: 37\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.23it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.17879818034405148 - train acc: 94.70118137595553\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 332.89it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.1682193907010463 - valid acc: 72.04968944099379\n",
      "Epoch: 38\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.28it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.16895561049270896 - train acc: 94.92703266157054\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 335.06it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.5150503504142048 - valid acc: 66.30434782608695\n",
      "Epoch: 39\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.25it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.18541198527679764 - train acc: 94.47533009034052\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 335.75it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 0.9979818316331985 - valid acc: 72.04968944099379\n",
      "Epoch: 40\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.31it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.1638103315617119 - train acc: 94.96177901320362\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 339.58it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.0245106155395591 - valid acc: 72.36024844720497\n",
      "Epoch: 41\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.20it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.188848340758981 - train acc: 94.21473245309242\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 345.66it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 0.9191674183366729 - valid acc: 73.75776397515527\n",
      "Epoch: 42\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.24it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.16383838622346936 - train acc: 94.78804725503822\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 352.74it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 0.9853141504756626 - valid acc: 74.37888198757764\n",
      "Epoch: 43\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.25it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.14602378655911824 - train acc: 95.53509381514941\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 353.78it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.0070697050000617 - valid acc: 73.91304347826086\n",
      "Epoch: 44\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.34it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.14144485267971815 - train acc: 95.81306462821404\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 344.62it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 0.9708263041820936 - valid acc: 72.98136645962732\n",
      "Epoch: 45\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.31it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.12955401965817592 - train acc: 96.47324530924253\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 342.68it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.0671623736352067 - valid acc: 73.91304347826086\n",
      "Epoch: 46\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.24it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.1283605791857289 - train acc: 96.14315496872828\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 333.40it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.1527585414824029 - valid acc: 72.67080745341616\n",
      "Epoch: 47\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.18it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.13114004243456617 - train acc: 96.02154273801251\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 354.46it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.4027554433768323 - valid acc: 70.65217391304348\n",
      "Epoch: 48\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.34it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.12210394511800572 - train acc: 96.49061848505906\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 349.07it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.3821919194015353 - valid acc: 72.20496894409938\n",
      "Epoch: 49\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.35it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.1210180384434718 - train acc: 96.21264767199445\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 327.17it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.3045416721539997 - valid acc: 70.80745341614907\n",
      "Epoch: 50\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.26it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.1009878127643183 - train acc: 97.09867963863795\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 341.30it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.0253988370612042 - valid acc: 72.98136645962732\n",
      "Epoch: 51\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.27it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.0984658070803521 - train acc: 97.28978457261988\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 333.62it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.0172266691558798 - valid acc: 74.06832298136646\n",
      "Epoch: 52\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.22it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.10168371967679961 - train acc: 97.01181375955524\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 336.87it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.0312753247526847 - valid acc: 73.75776397515527\n",
      "Epoch: 53\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.23it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.09997745933188074 - train acc: 97.20291869353717\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 329.81it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.062172014862646 - valid acc: 72.36024844720497\n",
      "Epoch: 54\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.26it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.10232201596857282 - train acc: 96.89020152883947\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 344.30it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.0369541717706436 - valid acc: 74.37888198757764\n",
      "Epoch: 55\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.25it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.1015262602771211 - train acc: 97.2550382209868\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 345.39it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.286235978202192 - valid acc: 66.92546583850931\n",
      "Epoch: 56\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.39it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.0812723184129296 - train acc: 97.75886031966643\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 335.12it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.27449595494706 - valid acc: 70.80745341614907\n",
      "Epoch: 57\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.21it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.08946045032863677 - train acc: 97.58512856150105\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 337.49it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.2483613728831309 - valid acc: 71.73913043478261\n",
      "Epoch: 58\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.34it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.08140527983013977 - train acc: 97.53300903405142\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 350.52it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.1321986917183897 - valid acc: 71.27329192546584\n",
      "Epoch: 59\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.34it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.1037741454172484 - train acc: 96.75121612230716\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 334.52it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.2585828019168057 - valid acc: 70.80745341614907\n",
      "Epoch: 60\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.29it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.09770843245650969 - train acc: 97.06393328700487\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 334.11it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.1375473184128524 - valid acc: 70.1863354037267\n",
      "Epoch: 61\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.34it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.09026158187413515 - train acc: 97.22029186935372\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 337.70it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 0.9951303100547169 - valid acc: 75.93167701863354\n",
      "Epoch: 62\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.26it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.09161932040630273 - train acc: 97.28978457261988\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 345.43it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.0966483617749878 - valid acc: 72.82608695652173\n",
      "Epoch: 63\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.23it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.0919865533974875 - train acc: 97.34190410006948\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 340.01it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.1104256477744212 - valid acc: 73.91304347826086\n",
      "Epoch: 64\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.21it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.10675307806040155 - train acc: 96.90757470465601\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 337.70it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.0865769653900903 - valid acc: 74.37888198757764\n",
      "Epoch: 65\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.29it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.10470491109998616 - train acc: 97.11605281445448\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 349.42it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.2868279825256652 - valid acc: 68.16770186335404\n",
      "Epoch: 66\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.35it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.08844111995074336 - train acc: 97.22029186935372\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 332.09it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.2826782779807795 - valid acc: 71.42857142857143\n",
      "Epoch: 67\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.27it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.06862876837157421 - train acc: 98.21056289089645\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 335.53it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.4625223761938546 - valid acc: 69.87577639751554\n",
      "Epoch: 68\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.22it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.0669737886403318 - train acc: 98.24530924252953\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 335.40it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.2732669920177726 - valid acc: 71.73913043478261\n",
      "Epoch: 69\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.23it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.07733675836996493 - train acc: 97.88047255038221\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 331.07it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.2524082838654254 - valid acc: 72.20496894409938\n",
      "Epoch: 70\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.34it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.07078051922697179 - train acc: 98.01945795691452\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 343.83it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.116340800104813 - valid acc: 73.29192546583852\n",
      "Epoch: 71\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.24it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.06825965192876549 - train acc: 97.94996525364836\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 339.36it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.111708802607358 - valid acc: 73.13664596273291\n",
      "Epoch: 72\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.27it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.07931613498471428 - train acc: 97.84572619874913\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 341.89it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.116333208623254 - valid acc: 74.06832298136646\n",
      "Epoch: 73\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.25it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.06366686912139034 - train acc: 98.33217512161224\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 334.20it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.099733335945977 - valid acc: 74.22360248447205\n",
      "Epoch: 74\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.25it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.08245896677325308 - train acc: 97.60250173731758\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 340.10it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.1164818937422514 - valid acc: 72.20496894409938\n",
      "Epoch: 75\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.32it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.07427614426250731 - train acc: 97.86309937456566\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 344.37it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.1817701505710787 - valid acc: 72.82608695652173\n",
      "Epoch: 76\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.22it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.053509638998553405 - train acc: 98.73175816539263\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 359.00it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.3970961082164188 - valid acc: 69.56521739130434\n",
      "Epoch: 77\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.27it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.04702142115380594 - train acc: 98.81862404447533\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 348.63it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.256554627470857 - valid acc: 71.8944099378882\n",
      "Epoch: 78\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.30it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.04776813924005619 - train acc: 98.73175816539263\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 336.56it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.500858809971492 - valid acc: 70.96273291925466\n",
      "Epoch: 79\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.26it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.07333766120387106 - train acc: 98.03683113273107\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 332.36it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.1644543255625308 - valid acc: 73.6024844720497\n",
      "Epoch: 80\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.27it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.035696237188030305 - train acc: 99.28769979152189\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 349.44it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.1668319409615635 - valid acc: 73.75776397515527\n",
      "Epoch: 81\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.25it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.04509534696689364 - train acc: 98.67963863794301\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 346.07it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.150634372428732 - valid acc: 73.13664596273291\n",
      "Epoch: 82\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.29it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.06814221754233098 - train acc: 98.07157748436414\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 335.58it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.22693291207556 - valid acc: 72.36024844720497\n",
      "Epoch: 83\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.34it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.09567397118769794 - train acc: 97.06393328700487\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 353.39it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.1166642759355694 - valid acc: 73.91304347826086\n",
      "Epoch: 84\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.34it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.09429373157954665 - train acc: 97.20291869353717\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 342.22it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.1440661236614613 - valid acc: 73.4472049689441\n",
      "Epoch: 85\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.27it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.07691857518430528 - train acc: 97.72411396803335\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 332.87it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.1896689755132148 - valid acc: 71.8944099378882\n",
      "Epoch: 86\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.34it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.07015662476441215 - train acc: 98.19318971507992\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 346.93it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.94441809729126 - valid acc: 62.732919254658384\n",
      "Epoch: 87\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.26it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.05674972487145296 - train acc: 98.33217512161224\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 338.26it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.3241261867702003 - valid acc: 69.87577639751554\n",
      "Epoch: 88\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.34it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.05411870618007411 - train acc: 98.76650451702571\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 334.10it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.4318781880353872 - valid acc: 71.8944099378882\n",
      "Epoch: 89\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.28it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.05512111675249173 - train acc: 98.59277275886032\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 339.58it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.2423844484343203 - valid acc: 72.67080745341616\n",
      "Epoch: 90\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.24it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.04593917141621672 - train acc: 98.59277275886032\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 331.30it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.2260108944972374 - valid acc: 72.67080745341616\n",
      "Epoch: 91\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.30it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.03956466188066374 - train acc: 99.06184850590688\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 347.39it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.2355990228500342 - valid acc: 73.91304347826086\n",
      "Epoch: 92\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.21it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.06015021010177137 - train acc: 98.34954829742877\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 329.48it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.2611433363500677 - valid acc: 72.36024844720497\n",
      "Epoch: 93\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.25it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.05032300215224897 - train acc: 98.74913134120918\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 336.56it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.2354756554714652 - valid acc: 73.13664596273291\n",
      "Epoch: 94\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.28it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.04922536943451236 - train acc: 98.50590687977763\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 337.48it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.2465283804614937 - valid acc: 73.6024844720497\n",
      "Epoch: 95\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.22it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.06871792613840844 - train acc: 98.03683113273107\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 333.95it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.3692932626210585 - valid acc: 69.72049689440993\n",
      "Epoch: 96\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.27it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.05403696255974326 - train acc: 98.59277275886032\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 337.01it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.633410301048647 - valid acc: 71.42857142857143\n",
      "Epoch: 97\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.20it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.04799698740139853 - train acc: 98.69701181375954\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 343.48it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.5226720733256416 - valid acc: 66.30434782608695\n",
      "Epoch: 98\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.22it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.04348499441126559 - train acc: 98.69701181375954\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 341.10it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.5898940387113303 - valid acc: 73.4472049689441\n",
      "Epoch: 99\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "180it [00:08, 21.23it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "train loss: 0.04844276675064762 - train acc: 98.47116052814454\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n",
      "644it [00:01, 352.03it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "valid loss: 1.2057398445076515 - valid acc: 74.84472049689441\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "for e in range(epoch):\n",
    "    model.train()\n",
    "    print(f\"Epoch: {e}\")\n",
    "    batch_cnt = 0\n",
    "    total_loss = 0\n",
    "    correct = 0\n",
    "    for batch, (train_sig, train_label) in tqdm(enumerate(traindl)):\n",
    "        batch_cnt = batch\n",
    "        train_sig = train_sig.to(device)\n",
    "        train_label = train_label.to(device)\n",
    "        \n",
    "        pred = model(train_sig)\n",
    "        loss = loss_fn(pred, train_label)\n",
    "        \n",
    "        optimizer.zero_grad()\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "        \n",
    "        scheduler.step()\n",
    "        \n",
    "        total_loss += loss.item()\n",
    "        correct += (pred.argmax(1) == train_label).type(torch.float).sum().item()\n",
    "    \n",
    "    total_loss /= batch_cnt\n",
    "    correct /= len(traindl.dataset)\n",
    "    \n",
    "    print(f\"train loss: {total_loss} - train acc: {100*correct}\")\n",
    "    \n",
    "    batch_cnt = 0\n",
    "    val_total_loss = 0\n",
    "    val_correct = 0\n",
    "    model.eval()\n",
    "    with torch.no_grad():\n",
    "        for batch, (valid_sig, valid_label) in tqdm(enumerate(validdl)):\n",
    "            batch_cnt = batch\n",
    "            valid_sig = valid_sig.to(device)\n",
    "            valid_label = valid_label.to(device)\n",
    "            \n",
    "            pred = model(valid_sig)\n",
    "            loss = loss_fn(pred, valid_label)\n",
    "            \n",
    "            val_total_loss += loss.item()\n",
    "            val_correct += (pred.argmax(1) == valid_label).type(torch.float).sum().item()\n",
    "    \n",
    "        val_total_loss /= batch_cnt\n",
    "        val_correct /= len(validdl.dataset)\n",
    "        \n",
    "        print(f\"valid loss: {val_total_loss} - valid acc: {100*val_correct}\")"
   ]
  }
 ],
 "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": "3.10.7"
  },
  "widgets": {
   "application/vnd.jupyter.widget-state+json": {
    "state": {},
    "version_major": 2,
    "version_minor": 0
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}