1-Multi-Class-Image-Classification-Fashion-MNIST.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Project Repository:** https://github.com/GokulKarthik/deep-learning-projects-pytorch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import time\n",
    "from tqdm.notebook import tqdm\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "import torch.optim as optim\n",
    "from torch.utils.data import Dataset, DataLoader\n",
    "from torch.utils.tensorboard import SummaryWriter\n",
    "from torchvision import datasets, transforms, utils\n",
    "from torchsummary import summary\n",
    "from sklearn.metrics import accuracy_score, confusion_matrix\n",
    "\n",
    "from PIL import Image"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "writer = SummaryWriter('runs/fashion-mnist-1')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. Load CSV files"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The csv files can be download from [this link](https://www.kaggle.com/zalando-research/fashionmnist)"
   ]
  },
  {
   "cell_type": "raw",
   "metadata": {},
   "source": [
    "\"\"\"\n",
    "#To load data from the package\n",
    "transform = transforms.Compose([\n",
    "    transforms.ToTensor(),\n",
    "    transforms.Normalize((0.5, ), (0.5, ))\n",
    "])\n",
    "\n",
    "train_set = datasets.FashionMNIST(root=\"data\", train=True, transform=transform, download=True)\n",
    "test_set = datasets.FashionMNIST(root=\"data\", train=False, transform=transform, download=True)\n",
    "\"\"\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "storage/data/fashion-mnist/fashion-mnist_train.csv storage/data/fashion-mnist/fashion-mnist_test.csv\n"
     ]
    }
   ],
   "source": [
    "data_path = os.path.join(\"storage\", \"data\", \"fashion-mnist\")\n",
    "train_path = os.path.join(data_path, \"fashion-mnist_train.csv\")\n",
    "test_path = os.path.join(data_path, \"fashion-mnist_test.csv\")\n",
    "print(train_path, test_path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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>label</th>\n",
       "      <th>pixel1</th>\n",
       "      <th>pixel2</th>\n",
       "      <th>pixel3</th>\n",
       "      <th>pixel4</th>\n",
       "      <th>pixel5</th>\n",
       "      <th>pixel6</th>\n",
       "      <th>pixel7</th>\n",
       "      <th>pixel8</th>\n",
       "      <th>pixel9</th>\n",
       "      <th>...</th>\n",
       "      <th>pixel775</th>\n",
       "      <th>pixel776</th>\n",
       "      <th>pixel777</th>\n",
       "      <th>pixel778</th>\n",
       "      <th>pixel779</th>\n",
       "      <th>pixel780</th>\n",
       "      <th>pixel781</th>\n",
       "      <th>pixel782</th>\n",
       "      <th>pixel783</th>\n",
       "      <th>pixel784</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>9</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>6</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>30</td>\n",
       "      <td>43</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 785 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   label  pixel1  pixel2  pixel3  pixel4  pixel5  pixel6  pixel7  pixel8  \\\n",
       "0      2       0       0       0       0       0       0       0       0   \n",
       "1      9       0       0       0       0       0       0       0       0   \n",
       "2      6       0       0       0       0       0       0       0       5   \n",
       "3      0       0       0       0       1       2       0       0       0   \n",
       "4      3       0       0       0       0       0       0       0       0   \n",
       "\n",
       "   pixel9  ...  pixel775  pixel776  pixel777  pixel778  pixel779  pixel780  \\\n",
       "0       0  ...         0         0         0         0         0         0   \n",
       "1       0  ...         0         0         0         0         0         0   \n",
       "2       0  ...         0         0         0        30        43         0   \n",
       "3       0  ...         3         0         0         0         0         1   \n",
       "4       0  ...         0         0         0         0         0         0   \n",
       "\n",
       "   pixel781  pixel782  pixel783  pixel784  \n",
       "0         0         0         0         0  \n",
       "1         0         0         0         0  \n",
       "2         0         0         0         0  \n",
       "3         0         0         0         0  \n",
       "4         0         0         0         0  \n",
       "\n",
       "[5 rows x 785 columns]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df = pd.read_csv(train_path) \n",
    "train_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "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>label</th>\n",
       "      <th>pixel1</th>\n",
       "      <th>pixel2</th>\n",
       "      <th>pixel3</th>\n",
       "      <th>pixel4</th>\n",
       "      <th>pixel5</th>\n",
       "      <th>pixel6</th>\n",
       "      <th>pixel7</th>\n",
       "      <th>pixel8</th>\n",
       "      <th>pixel9</th>\n",
       "      <th>...</th>\n",
       "      <th>pixel775</th>\n",
       "      <th>pixel776</th>\n",
       "      <th>pixel777</th>\n",
       "      <th>pixel778</th>\n",
       "      <th>pixel779</th>\n",
       "      <th>pixel780</th>\n",
       "      <th>pixel781</th>\n",
       "      <th>pixel782</th>\n",
       "      <th>pixel783</th>\n",
       "      <th>pixel784</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>9</td>\n",
       "      <td>8</td>\n",
       "      <td>...</td>\n",
       "      <td>103</td>\n",
       "      <td>87</td>\n",
       "      <td>56</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>34</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>14</td>\n",
       "      <td>53</td>\n",
       "      <td>99</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>63</td>\n",
       "      <td>53</td>\n",
       "      <td>31</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>137</td>\n",
       "      <td>126</td>\n",
       "      <td>140</td>\n",
       "      <td>0</td>\n",
       "      <td>133</td>\n",
       "      <td>224</td>\n",
       "      <td>222</td>\n",
       "      <td>56</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 785 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   label  pixel1  pixel2  pixel3  pixel4  pixel5  pixel6  pixel7  pixel8  \\\n",
       "0      0       0       0       0       0       0       0       0       9   \n",
       "1      1       0       0       0       0       0       0       0       0   \n",
       "2      2       0       0       0       0       0       0      14      53   \n",
       "3      2       0       0       0       0       0       0       0       0   \n",
       "4      3       0       0       0       0       0       0       0       0   \n",
       "\n",
       "   pixel9  ...  pixel775  pixel776  pixel777  pixel778  pixel779  pixel780  \\\n",
       "0       8  ...       103        87        56         0         0         0   \n",
       "1       0  ...        34         0         0         0         0         0   \n",
       "2      99  ...         0         0         0         0        63        53   \n",
       "3       0  ...       137       126       140         0       133       224   \n",
       "4       0  ...         0         0         0         0         0         0   \n",
       "\n",
       "   pixel781  pixel782  pixel783  pixel784  \n",
       "0         0         0         0         0  \n",
       "1         0         0         0         0  \n",
       "2        31         0         0         0  \n",
       "3       222        56         0         0  \n",
       "4         0         0         0         0  \n",
       "\n",
       "[5 rows x 785 columns]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_df = pd.read_csv(test_path)\n",
    "test_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(60000, 785) (10000, 785)\n",
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 60000 entries, 0 to 59999\n",
      "Columns: 785 entries, label to pixel784\n",
      "dtypes: int64(785)\n",
      "memory usage: 359.3 MB\n",
      "None\n",
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 10000 entries, 0 to 9999\n",
      "Columns: 785 entries, label to pixel784\n",
      "dtypes: int64(785)\n",
      "memory usage: 59.9 MB\n",
      "None\n"
     ]
    }
   ],
   "source": [
    "print(train_df.shape, test_df.shape)\n",
    "print(train_df.info())\n",
    "print(test_df.info())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "255"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df.iloc[0, 1:].max()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "9    6000\n",
       "8    6000\n",
       "7    6000\n",
       "6    6000\n",
       "5    6000\n",
       "4    6000\n",
       "3    6000\n",
       "2    6000\n",
       "1    6000\n",
       "0    6000\n",
       "Name: label, dtype: int64"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_df['label'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "7    1000\n",
       "6    1000\n",
       "5    1000\n",
       "4    1000\n",
       "3    1000\n",
       "2    1000\n",
       "9    1000\n",
       "1    1000\n",
       "8    1000\n",
       "0    1000\n",
       "Name: label, dtype: int64"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_df['label'].value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. Format into images"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Loading all the data in the CSV file is not memory efficient. Also, We cannot understand the data in raw format. Hence, I am going to format this file into images in the following format.\n",
    "* train/0/xxx.png\n",
    "* train/1/xxx.png\n",
    "* train/2/xx.png\n",
    "* ..."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "storage/data/fashion-mnist/train storage/data/fashion-mnist/test\n"
     ]
    }
   ],
   "source": [
    "train_path = os.path.join(data_path, \"train\")\n",
    "test_path = os.path.join(data_path, \"test\")\n",
    "print(train_path, test_path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "for path in [train_path, test_path]:\n",
    "    if not os.path.exists(path):\n",
    "        os.mkdir(path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[2, 9, 6, 0, 3, 4, 5, 8, 7, 1]\n"
     ]
    }
   ],
   "source": [
    "class_labels = train_df['label'].unique().tolist()\n",
    "print(class_labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "for path in [train_path, test_path]:\n",
    "    for class_label in class_labels:\n",
    "        class_path = os.path.join(path, str(class_label))\n",
    "        if not os.path.exists(class_path):\n",
    "            os.mkdir(class_path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "d3e1275235574d29a940a3c9c068f2c4",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=60000.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "for row_id, row in tqdm(train_df.iterrows(), total=len(train_df)):\n",
    "    label = row['label']\n",
    "    image_path = os.path.join(data_path, \"train\", str(label), \"{}.png\".format(row_id))\n",
    "    pixels = row.iloc[1:].values.astype(\"uint8\")\n",
    "    pixels = pixels.reshape(28, 28)\n",
    "    image = Image.fromarray(pixels)\n",
    "    image.save(image_path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "81cac8f7953d45c1995f3530f53bc243",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=10000.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "for row_id, row in tqdm(test_df.iterrows(), total=len(test_df)):\n",
    "    label = row['label']\n",
    "    image_path = os.path.join(data_path, \"test\", str(label), \"{}.png\".format(row_id))\n",
    "    pixels = row.iloc[1:].values.astype(\"uint8\")\n",
    "    pixels = pixels.reshape(28, 28)\n",
    "    image = Image.fromarray(pixels)\n",
    "    image.save(image_path)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Note**: `del dataframe` does not release memory. So restart the notebook and start executimg from the next section after generating images."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3. Define dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "storage/data/fashion-mnist/train storage/data/fashion-mnist/test\n"
     ]
    }
   ],
   "source": [
    "data_path = os.path.join(\"storage\", \"data\", \"fashion-mnist\")\n",
    "train_path = os.path.join(data_path, \"train\")\n",
    "test_path = os.path.join(data_path, \"test\")\n",
    "print(train_path, test_path)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "`torchvision.transforms.ToTensor()` squishes the value range from [0, 255] to [0., 1.]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "mean = 0.5\n",
    "sd = 0.5\n",
    "transform = transforms.Compose([\n",
    "    transforms.Grayscale(),\n",
    "    transforms.ToTensor(),\n",
    "    transforms.Normalize((mean, ), (sd, ))\n",
    "])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We can use `torchvision.datasets.ImageFolder()` to define the dataset instead of writing a dataset class from scratch."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_set = datasets.ImageFolder(root=train_path, transform=transform)\n",
    "test_set = datasets.ImageFolder(root=test_path, transform=transform)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 4. Define dataloader"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_batch_size = 64\n",
    "test_batch_size = 4"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "8\n"
     ]
    }
   ],
   "source": [
    "num_cpus = os.cpu_count()\n",
    "print(num_cpus)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_loader = DataLoader(train_set, batch_size=train_batch_size, shuffle=True, num_workers=num_cpus) \n",
    "test_loader = DataLoader(test_set, batch_size=4, shuffle=test_batch_size, num_workers=num_cpus)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "938\n"
     ]
    }
   ],
   "source": [
    "len_train_loader = len(train_loader)\n",
    "print(len_train_loader)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 5. Define utilities"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "label_to_name = {\n",
    "    \"0\": \"t-shirt\",\n",
    "    \"1\": \"trouser\",\n",
    "    \"2\": \"pullover\",\n",
    "    \"3\": \"dress\",\n",
    "    \"4\": \"coat\",\n",
    "    \"5\": \"sandal\", \n",
    "    \"6\": \"shirt\",\n",
    "    \"7\": \"sneaker\",\n",
    "    \"8\": \"bag\",\n",
    "    \"9\": \"ankle_boot\"\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'t-shirt': '0', 'trouser': '1', 'pullover': '2', 'dress': '3', 'coat': '4', 'sandal': '5', 'shirt': '6', 'sneaker': '7', 'bag': '8', 'ankle_boot': '9'}\n"
     ]
    }
   ],
   "source": [
    "name_to_label = {v: k for k, v in label_to_name.items()}\n",
    "print(name_to_label)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_image(image_tensor, title=\"\"):\n",
    "    \"\"\"\n",
    "    Arguments:\n",
    "    image_tensor -- tensor of Size([1, n_h, n_w]) / Size([3, n_h+4, n_w*m+4])\n",
    "    \"\"\"\n",
    "    # image_tensor.squeeze() -> doesn't work for grids of Size([3, n_h+4, n_c*m+4])\n",
    "    image_tensor = image_tensor.mean(dim = 0) # Size([28, 28])\n",
    "    image_tensor = (image_tensor * sd) + mean\n",
    "    image_tensor = image_tensor * 255 # not necessary\n",
    "    image_array = image_tensor.numpy()\n",
    "    plt.figure(figsize=(20, 20))\n",
    "    plt.imshow(image_array, cmap='Greys')\n",
    "    plt.title(title)\n",
    "    return plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x1440 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "train_iter = iter(train_loader)\n",
    "X, Y = train_iter.next()\n",
    "m = min(16, len(train_loader))\n",
    "image_grid = utils.make_grid(X[:m])\n",
    "names = [label_to_name[str(label.item())] for label in Y[:m]]\n",
    "plot_image(image_grid, title=names)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 6. Define model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "cuda:0\n"
     ]
    }
   ],
   "source": [
    "device = \"cuda:0\" if torch.cuda.is_available() else \"cpu\"\n",
    "device = torch.device(device)\n",
    "print(device)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The following architecture is modified from the LeNet"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "class Model(nn.Module):\n",
    "    \n",
    "    def __init__(self):\n",
    "        super(Model, self).__init__()\n",
    "        self.conv1 = nn.Conv2d(in_channels=1, out_channels=6, kernel_size=5, stride=1, padding=0) # [m, 6, 24, 24]\n",
    "        self.pool1 = nn.MaxPool2d(kernel_size=2, stride=2) # [m, 6, 12, 12]\n",
    "        self.conv2 = nn.Conv2d(in_channels=6, out_channels=16, kernel_size=5, stride=1, padding=0) # [m, 16, 8, 8]\n",
    "        self.pool2 = nn.MaxPool2d(kernel_size=2, stride=2) # [m, 16, 4, 4]\n",
    "        self.fc3 = nn.Linear(in_features=16*4*4, out_features=120) # [m, 120]\n",
    "        self.fc4 = nn.Linear(in_features=120, out_features=84) # [m, 84]\n",
    "        self.fc5 = nn.Linear(in_features=84, out_features=10) # [m, 10]\n",
    "        \n",
    "    def forward(self, X):\n",
    "        X = self.pool1(F.relu(self.conv1(X)))\n",
    "        X = self.pool2(F.relu(self.conv2(X)))\n",
    "        X = X.view(-1, 16*4*4)\n",
    "        X = F.relu(self.fc3(X))\n",
    "        X = F.relu(self.fc4(X))\n",
    "        X = self.fc5(X) \n",
    "        return X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "DataParallel(\n",
      "  (module): Model(\n",
      "    (conv1): Conv2d(1, 6, kernel_size=(5, 5), stride=(1, 1))\n",
      "    (pool1): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
      "    (conv2): Conv2d(6, 16, kernel_size=(5, 5), stride=(1, 1))\n",
      "    (pool2): MaxPool2d(kernel_size=2, stride=2, padding=0, dilation=1, ceil_mode=False)\n",
      "    (fc3): Linear(in_features=256, out_features=120, bias=True)\n",
      "    (fc4): Linear(in_features=120, out_features=84, bias=True)\n",
      "    (fc5): Linear(in_features=84, out_features=10, bias=True)\n",
      "  )\n",
      ")\n"
     ]
    }
   ],
   "source": [
    "model = Model()\n",
    "model = nn.DataParallel(model)\n",
    "model = model.to(device)\n",
    "print(model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "#list(model.parameters())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------------------------------------------------\n",
      "        Layer (type)               Output Shape         Param #\n",
      "================================================================\n",
      "            Conv2d-1            [-1, 6, 24, 24]             156\n",
      "         MaxPool2d-2            [-1, 6, 12, 12]               0\n",
      "            Conv2d-3             [-1, 16, 8, 8]           2,416\n",
      "         MaxPool2d-4             [-1, 16, 4, 4]               0\n",
      "            Linear-5                  [-1, 120]          30,840\n",
      "            Linear-6                   [-1, 84]          10,164\n",
      "            Linear-7                   [-1, 10]             850\n",
      "             Model-8                   [-1, 10]               0\n",
      "================================================================\n",
      "Total params: 44,426\n",
      "Trainable params: 44,426\n",
      "Non-trainable params: 0\n",
      "----------------------------------------------------------------\n",
      "Input size (MB): 0.00\n",
      "Forward/backward pass size (MB): 0.04\n",
      "Params size (MB): 0.17\n",
      "Estimated Total Size (MB): 0.22\n",
      "----------------------------------------------------------------\n"
     ]
    }
   ],
   "source": [
    "summary(model, input_size=(1, 28, 28))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "writer.add_graph(model, X)\n",
    "writer.close()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Note:** Use `tensorboard --logdir=runs` command in the Terminal to view the TensorBoard"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 7. Set optimizer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3752\n"
     ]
    }
   ],
   "source": [
    "lr = 0.001\n",
    "step_size = len(train_loader) * 4\n",
    "gamma = 0.95\n",
    "print(step_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "criterion = nn.CrossEntropyLoss(reduction=\"mean\")\n",
    "optimizer = optim.Adam(model.parameters(), lr=lr)\n",
    "lr_scheduler = optim.lr_scheduler.StepLR(optimizer=optimizer, step_size=step_size, gamma=gamma)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 8. Train model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "epochs = 20\n",
    "print_every_n_epochs = 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "epoch_losses = []\n",
    "epoch_lrs = []\n",
    "iteration_losses = []\n",
    "iteration_lrs = []\n",
    "\n",
    "for epoch in tqdm(range(1, epochs+1)):\n",
    "    epoch_loss = 0\n",
    "    epoch_lr = 0\n",
    "    \n",
    "    for X, Y in tqdm(train_loader, desc=\"Epoch-{}\".format(epoch)):\n",
    "        X, Y = X.to(device), Y.to(device)\n",
    "        \n",
    "        optimizer.zero_grad()\n",
    "        Y_pred_logits = model(X)\n",
    "        loss = criterion(Y_pred_logits, Y)\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "        lr_scheduler.step()\n",
    "        \n",
    "        iteration_losses.append(loss.item())\n",
    "        iteration_lrs.append(lr_scheduler.get_lr()[0])\n",
    "        epoch_loss += loss.item()\n",
    "        epoch_lr += lr_scheduler.get_lr()[0]\n",
    "        \n",
    "    epoch_loss /= len(train_loader)\n",
    "    epoch_lr /= len(train_loader)\n",
    "    epoch_losses.append(epoch_loss)\n",
    "    epoch_lrs.append(epoch_lr)\n",
    "    \n",
    "    if epoch % print_every_n_epochs == 0:    \n",
    "        message = \"Epoch:{}    Loss:{}    LR:{}\".format(epoch, epoch_loss, epoch_lr)\n",
    "        print(message)\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x576 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax1, ax2) = plt.subplots(2, 1, sharex=True, figsize=(15, 8))\n",
    "ax1.plot(epoch_losses, marker=\"o\", markersize=5)\n",
    "ax1.set_title(\"Loss\")\n",
    "ax2.plot(epoch_lrs, marker=\"o\", markersize=5)\n",
    "ax2.set_title(\"LR\")\n",
    "plt.xlabel(\"Epochs\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x576 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax1, ax2) = plt.subplots(2, 1, sharex=True, figsize=(15, 8))\n",
    "ax1.plot(iteration_losses[::100])\n",
    "ax1.set_title(\"Loss\")\n",
    "ax2.plot(iteration_lrs[::100])\n",
    "ax2.set_title(\"LR\")\n",
    "plt.xlabel(\"Iterations\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "window = 100\n",
    "pd.Series(iteration_losses).rolling(window=window).mean().iloc[window-1:].plot()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "path = os.path.join(\"storage\", \"models\", \"fashion-mnist\", \"classifier.pth\")\n",
    "torch.save(model.state_dict(), path)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 9. Test model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<All keys matched successfully>"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "path = os.path.join(\"storage\", \"models\", \"fashion-mnist\", \"classifier.pth\")\n",
    "model = Model()\n",
    "model = nn.DataParallel(model)\n",
    "model.load_state_dict(torch.load(path))\n",
    "#model = model.to(\"cpu\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "True\n",
      "True\n",
      "True\n",
      "True\n",
      "True\n",
      "True\n",
      "True\n",
      "True\n",
      "True\n",
      "True\n"
     ]
    }
   ],
   "source": [
    "for p in model.parameters():\n",
    "    print(p.is_cuda)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "b513bff3298046fba22fd1ac76049d06",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=938.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "with torch.no_grad():\n",
    "    Y_train, Y_pred_train = [], []\n",
    "    for X_mb, Y_mb in tqdm(train_loader):\n",
    "        out = model(X_mb)\n",
    "        _, Y_pred_mb = torch.max(out, 1)\n",
    "        Y_train.extend(Y_mb.numpy().tolist())\n",
    "        Y_pred_train.extend(Y_pred_mb.cpu().numpy().tolist())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "4bf43876d0ea44bbaf53783ab760da97",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=0.0, max=2500.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "with torch.no_grad():\n",
    "    Y_test, Y_pred_test = [], []\n",
    "    for X_mb, Y_mb in tqdm(test_loader):\n",
    "        out = model(X_mb)\n",
    "        _, Y_pred_mb = torch.max(out, 1)\n",
    "        Y_test.extend(Y_mb.numpy().tolist())\n",
    "        Y_pred_test.extend(Y_pred_mb.cpu().numpy().tolist())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Accuracy: 0.9404166666666667\n",
      "Test Accuracy: 0.8984\n"
     ]
    }
   ],
   "source": [
    "train_accuracy = accuracy_score(Y_train, Y_pred_train)\n",
    "test_accuracy = accuracy_score(Y_test, Y_pred_test)\n",
    "print(\"Train Accuracy: {}\".format(train_accuracy))\n",
    "print(\"Test Accuracy: {}\".format(test_accuracy))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "labels = [label_to_name[str(i)] for i in range(10)]\n",
    "c_mat_train = confusion_matrix(Y_train, Y_pred_train)\n",
    "c_mat_test = confusion_matrix(Y_test, Y_pred_test)\n",
    "fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(15, 5))\n",
    "sns.heatmap(c_mat_train, annot=True, fmt='d', ax=ax1, xticklabels=labels, yticklabels=labels)\n",
    "ax1.set_title('Train Data')\n",
    "ax1.set_xlabel('Predicted')\n",
    "ax1.set_ylabel('Actual')\n",
    "sns.heatmap(c_mat_test, annot=True, fmt='d', ax=ax2, xticklabels=labels, yticklabels=labels)\n",
    "ax2.set_title('Test Data')\n",
    "ax2.set_xlabel('Predicted')\n",
    "ax2.set_ylabel('Actual')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "def compute_accuracies(c_mat):\n",
    "    accuracies = c_mat.astype('float') / c_mat.sum(axis=1)\n",
    "    accuracies = accuracies.diagonal()\n",
    "    accuracies = {k:v for k, v in zip(labels, accuracies)}\n",
    "    return accuracies"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'t-shirt': 0.9366666666666666,\n",
       " 'trouser': 0.9981666666666666,\n",
       " 'pullover': 0.8301666666666667,\n",
       " 'dress': 0.93,\n",
       " 'coat': 0.966,\n",
       " 'sandal': 0.996,\n",
       " 'shirt': 0.7988333333333333,\n",
       " 'sneaker': 0.9765,\n",
       " 'bag': 0.9926666666666667,\n",
       " 'ankle_boot': 0.9791666666666666}"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "compute_accuracies(c_mat_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'t-shirt': 0.865,\n",
       " 'trouser': 0.989,\n",
       " 'pullover': 0.759,\n",
       " 'dress': 0.892,\n",
       " 'coat': 0.929,\n",
       " 'sandal': 0.968,\n",
       " 'shirt': 0.7,\n",
       " 'sneaker': 0.943,\n",
       " 'bag': 0.972,\n",
       " 'ankle_boot': 0.967}"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "compute_accuracies(c_mat_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}