Train_using_embeddings_aug_b0.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "aaee65d4-fdcd-492e-8a12-24a9f6e6e9af",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Goal : To classify the type of ovarian cancer from microscopy scans of biopsy samples.\n",
    "\n",
    "# Search SoTA model on Kaggle Approach, Papers with code - Medical Data, Graph Transformer, GNN, Global and Local Analysis\n",
    "\n",
    "# You should manage, including differences in image dimensions, quality, slide staining techniques, and more\n",
    "\n",
    "# There are 2 types of images - TMA and WSI (25 images are TMA and 513(rest) are WSI)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "9d1971b2-09c3-4a8d-b158-355c9772dfa9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'/projectnb/cs640grp/students/avarshn/.conda/envs/ocean'"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import sys\n",
    "sys.prefix"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "48222a33-507b-4f98-8df3-8538ed479f1c",
   "metadata": {},
   "source": [
    "# Check if GPU is available"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "143df7b3-2bc3-4651-b9a8-ed3f5a8108be",
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "c193c329-3591-4572-9640-35b274bf754e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "torch.cuda.is_available()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "3c04ae72-d26a-4c3c-8415-ba36cb557c94",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "/bin/bash: nvidia: command not found\n"
     ]
    }
   ],
   "source": [
    "!nvidia smi\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "150de13e-2b4e-499e-9092-e541baa686b8",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "02:00.0 3D controller: NVIDIA Corporation GP100GL [Tesla P100 PCIe 12GB] (rev a1)\n",
      "82:00.0 3D controller: NVIDIA Corporation GP100GL [Tesla P100 PCIe 12GB] (rev a1)\n"
     ]
    }
   ],
   "source": [
    "!lspci | grep -i nvidia"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "75827adc-9618-4d47-8fdc-44dbdc43cbe8",
   "metadata": {},
   "source": [
    "# Import Libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "b7db4971-169f-47a7-9938-cc2d20e23b09",
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "from time import time\n",
    "from collections import defaultdict\n",
    "from tqdm import tqdm\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from sklearn.utils.class_weight import compute_class_weight\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import accuracy_score, f1_score, confusion_matrix, classification_report\n",
    "\n",
    "from PIL import Image\n",
    "import cv2\n",
    "\n",
    "import torch.nn as nn\n",
    "from torch.utils.data import Dataset, DataLoader\n",
    "from torch.optim import AdamW\n",
    "\n",
    "from torchsummary import summary\n",
    "\n",
    "import torchvision\n",
    "from torchvision import transforms, utils\n",
    "\n",
    "from torchvision.models import efficientnet_b0, EfficientNet_B0_Weights\n",
    "from torchvision.models import resnet18, ResNet18_Weights"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a5c70d5b-6ba8-4faf-a3ff-5e37b6a0c467",
   "metadata": {},
   "source": [
    "# Configurations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "13ca183a-b75e-4d06-ae54-1c3af198cae6",
   "metadata": {},
   "outputs": [],
   "source": [
    "CONFIG = {\n",
    "    \"model_name\" : \"efficientnet_b0\"\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "c572c798-bfbf-49bb-a47e-95935ba372dd",
   "metadata": {},
   "outputs": [],
   "source": [
    "if CONFIG[\"model_name\"] == \"efficientnet_b3\":\n",
    "\n",
    "    CONFIG.update({\n",
    "            \"seed\": 42,\n",
    "            \"num_classes\": 5,\n",
    "            \"device\": torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\"),\n",
    "            \n",
    "            # \"valid_batch_size\": 32,\n",
    "            \"embedding_save_folder\" : \"./models/efficientnet_b0/embeddings/\",\n",
    "            \"feature_input_size\" : 1536,\n",
    "            \"model_save_folder\" : \"./models/efficientnet_b0/model_checkpoints/\",\n",
    "        \n",
    "            \"batch_size\" : 8,\n",
    "            \"epochs\" : 50,\n",
    "            \"learning_rate\" : 1e-5,\n",
    "            \"weight_decay\" : 1e-3\n",
    "    })"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "id": "4dcff8b2-936b-458c-a5f1-b7b711030b2b",
   "metadata": {},
   "outputs": [],
   "source": [
    "if CONFIG[\"model_name\"] == \"efficientnet_b0\":\n",
    "\n",
    "    CONFIG.update({\n",
    "            \"seed\": 42,\n",
    "            \"num_classes\": 5,\n",
    "            \"device\": torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\"),\n",
    "            \n",
    "            # \"valid_batch_size\": 32,\n",
    "            \"embedding_save_folder\" : \"./models/efficientnet_b0/embeddings/\",\n",
    "            \"feature_input_size\" : 1280,\n",
    "            \"model_save_folder\" : \"./models/efficientnet_b0/model_checkpoints/\",\n",
    "        \n",
    "            \"batch_size\" : 8,\n",
    "            \"epochs\" : 40,\n",
    "            \"learning_rate\" : 1e-5,\n",
    "            \"weight_decay\" : 1e-2\n",
    "    })"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "c49af72f-38c9-44c7-b1a5-60dd6203ba46",
   "metadata": {},
   "outputs": [],
   "source": [
    "if CONFIG[\"model_name\"] == \"resnet18\":\n",
    "\n",
    "    CONFIG.update({\n",
    "            \"seed\": 42,\n",
    "            \"num_classes\": 5,\n",
    "            \"device\": torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\"),\n",
    "            \n",
    "            # \"valid_batch_size\": 32,\n",
    "            \"embedding_save_folder\" : \"./models/resnet18/embeddings/\",\n",
    "            \"feature_input_size\" : 512,\n",
    "            \"model_save_folder\" : \"./models/resnet18/model_checkpoints/\",\n",
    "        \n",
    "            \"batch_size\" : 32,\n",
    "            \"epochs\" : 50,\n",
    "            \"learning_rate\" : 1e-4,\n",
    "            \"weight_decay\" : 1e-3\n",
    "    })"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "d5fbe965-3c93-4b23-bb36-f4f56aca0000",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'model_name': 'efficientnet_b0',\n",
       " 'seed': 42,\n",
       " 'num_classes': 5,\n",
       " 'device': device(type='cuda', index=0),\n",
       " 'embedding_save_folder': './models/efficientnet_b0/embeddings/',\n",
       " 'feature_input_size': 1280,\n",
       " 'model_save_folder': './models/efficientnet_b0/model_checkpoints/',\n",
       " 'batch_size': 8,\n",
       " 'epochs': 20,\n",
       " 'learning_rate': 1e-05,\n",
       " 'weight_decay': 0.01}"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "CONFIG"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "59ef42e6-00e5-40ff-a2b9-9e44b718dcee",
   "metadata": {},
   "outputs": [],
   "source": [
    "def set_seed(seed=42):\n",
    "    '''Sets the seed of the entire notebook so results are the same every time we run.\n",
    "    This is for REPRODUCIBILITY.'''\n",
    "    np.random.seed(seed)\n",
    "    torch.manual_seed(seed)\n",
    "    torch.cuda.manual_seed(seed)\n",
    "    # When running on the CuDNN backend, two further options must be set\n",
    "    torch.backends.cudnn.deterministic = True\n",
    "    torch.backends.cudnn.benchmark = False\n",
    "    # Set a fixed value for the hash seed\n",
    "    os.environ['PYTHONHASHSEED'] = str(seed)\n",
    "    \n",
    "set_seed(CONFIG['seed'])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9b28ab07-67fb-4555-9187-88e860e73c34",
   "metadata": {},
   "source": [
    "# Merge Numpy arrays"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "859bccec-19a7-4315-9448-32e9edaa9b1c",
   "metadata": {},
   "outputs": [],
   "source": [
    "embeddings_arr_list = []\n",
    "labels_arr_list = []\n",
    "\n",
    "for file in os.listdir(CONFIG[\"embedding_save_folder\"]):\n",
    "    if file.endswith(\".npz\"):\n",
    "        if \"train_embeddings_and_labels_\" in file:\n",
    "            np_arr = np.load(CONFIG[\"embedding_save_folder\"] + file)\n",
    "            embeddings_arr_list.append(np_arr[\"embeddings\"])\n",
    "            labels_arr_list.append(np_arr[\"labels\"])\n",
    "            # labels_arr = np.concatenate((labels_arr, np_arr[\"labels\"]), axis = 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "9c04dbc4-4ea3-436a-8f6e-6839333c4503",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1505, 1280)"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "all_embeddings = np.concatenate(embeddings_arr_list)\n",
    "all_embeddings.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "4d2675b4-3009-47a7-bf8d-1f87046b58c8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1505,)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "all_labels = np.concatenate(labels_arr_list)\n",
    "all_labels.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "df552edb-4d06-45ca-8fcb-95f99079fe1d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([0, 1, 2, 3, 4]), array([329, 299, 285, 293, 299]))"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Balanced Dataset with Augmentation\n",
    "np.unique(all_labels, return_counts=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "7c39fe9e-d3ec-4df1-aee3-b8c790f94faa",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "./models/efficientnet_b0/embeddings/efficientnet_b0_train_embeddings_and_labels.npz\n"
     ]
    }
   ],
   "source": [
    "save_location = f'{CONFIG[\"embedding_save_folder\"]}{CONFIG[\"model_name\"]}_train_embeddings_and_labels.npz'\n",
    "print(save_location)\n",
    "\n",
    "np.savez(save_location, embeddings = all_embeddings, labels = all_labels)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2f3350a3-92d5-4526-8c18-60568057476c",
   "metadata": {},
   "source": [
    "# Custom Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "02dd6b7a-b315-49e5-ab7d-50141f6a92c5",
   "metadata": {},
   "outputs": [],
   "source": [
    "Image.MAX_IMAGE_PIXELS = None"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "002b334d-0f52-4bde-828c-eddcbb34a033",
   "metadata": {},
   "outputs": [],
   "source": [
    "class ocean_embedding_dataset(Dataset):\n",
    "    \n",
    "    def __init__(self, numpy_file_name):\n",
    "        \n",
    "        self.ocean_np_arr = np.load(numpy_file_name)\n",
    "\n",
    "    def __len__(self):\n",
    "        return self.ocean_np_arr[\"labels\"].shape[0]\n",
    "\n",
    "    def __getitem__(self, idx):\n",
    "        if torch.is_tensor(idx):\n",
    "            idx = idx.tolist()\n",
    "\n",
    "        image = self.ocean_np_arr[\"embeddings\"][idx]\n",
    "        image = torch.tensor(image)\n",
    "        \n",
    "        label = self.ocean_np_arr[\"labels\"][idx]\n",
    "\n",
    "        return image, label\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "33858d05-4c7f-4766-8cb5-84cc5b1dc770",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "01054b83-dfda-4717-a6b5-64608e69e9d5",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_np_arr_path = f'{CONFIG[\"embedding_save_folder\"]}{CONFIG[\"model_name\"]}_train_embeddings_and_labels.npz'\n",
    "test_np_arr_path = f'{CONFIG[\"embedding_save_folder\"]}{CONFIG[\"model_name\"]}_test_embeddings_and_labels.npz'\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "27eba013-ad6c-4dc8-b686-e7e327206c69",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_dataset = ocean_embedding_dataset(train_np_arr_path)\n",
    "\n",
    "test_dataset = ocean_embedding_dataset(test_np_arr_path)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "18e15d1c-a392-41a3-a03f-5642fcdbfb96",
   "metadata": {},
   "source": [
    "## Testing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "50fbd20d-5ad0-44d1-8c8a-5bd32dcbc02e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1505"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_dataset.__len__()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "86204af9-161a-488c-b890-b8f11b17911b",
   "metadata": {},
   "outputs": [],
   "source": [
    "a,b = train_dataset.__getitem__(torch.tensor(10))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "52061b6a-000b-4ff2-8595-e37f9007f17f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([ 0.0052, -0.0422,  0.2885,  ..., -0.1830, -0.1882, -0.1551])"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "a"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "bd5247e2-7402-4bdd-b77d-f5d75b2ce125",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "3"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "b"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "b46cb26c-214f-4d06-a8c4-905d0ec08154",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([4, 3, 2, ..., 1, 1, 4])"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_dataset.ocean_np_arr[\"labels\"]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "20401933-47b3-404e-ba4e-be65e40f6ae6",
   "metadata": {},
   "source": [
    "# Class Weights - From Training Data since the dataset is imbalanced"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2d5ad94e-b4fd-4454-8b6b-f8a607065cc8",
   "metadata": {},
   "source": [
    "## Not using Class Weights in nn.CrossEntropyLoss() since we have balanced data while Augmentation itself"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "31f40cfe-c181-440a-8f3e-d440782b30c3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([4, 3, 2, ..., 1, 1, 4])"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train_dataset.ocean_np_arr[\"labels\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6e8ff4c7-62e4-4654-a53b-209323c92e3f",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "75b28ad6-975a-4fc2-a19b-73bde1c09870",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "2b246783-1d97-4989-ac02-f7e263919015",
   "metadata": {},
   "source": [
    "* Class weight Formula -\n",
    "  \n",
    "class_weight_of_that_class = $\\frac{\\text{Total samples}}{\\text{Total classes * samples of that class}}$"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "d3548690-c5de-462c-b869-e2ed6836beb5",
   "metadata": {},
   "outputs": [],
   "source": [
    "# NumPy array of class labels - labels_arr\n",
    "\n",
    "labels_arr = train_dataset.ocean_np_arr[\"labels\"].copy()\n",
    "\n",
    "class_weights = compute_class_weight('balanced',classes = np.unique(labels_arr), y = labels_arr)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "3c848bee-deca-4c3d-ac8f-b6b0c58ae80e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0.91489362, 1.00668896, 1.05614035, 1.02730375, 1.00668896])"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "class_weights"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "42486cca-117d-4dcb-9a8b-a2a118939821",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([0, 1, 2, 3, 4]), array([329, 299, 285, 293, 299]))"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# label_mapper = {'HGSC' : 0, 'EC' : 1, 'MC' : 2, 'CC' : 3, 'LGSC' : 4}\n",
    "# Train Label Distribution\n",
    "np.unique(labels_arr, return_counts = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "4b19a847-2296-4df0-9897-8e5a295e6c6b",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([0.9149, 1.0067, 1.0561, 1.0273, 1.0067], dtype=torch.float64)"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "class_weights_tensor = torch.tensor(class_weights)\n",
    "class_weights_tensor"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d4d58fe0-b89b-44cb-847d-8085557d6e5b",
   "metadata": {},
   "source": [
    "# DataLoader"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b05e9ade-27f6-4b02-af47-feac75b859e6",
   "metadata": {},
   "source": [
    "* The EOFError you encountered in your DataLoader worker process could be related to using multiple workers (num_workers = 2)\n",
    "\n",
    "\n",
    "1) Concurrency Issues:\n",
    "\n",
    "    * When using multiple worker processes (num_workers > 0), there is potential for concurrency issues, especially when reading data from a file or other data sources. If your dataset is not designed to handle concurrent access properly, it can result in unexpected errors like EOFError.\n",
    "  \n",
    "2) Use torch.utils.data.SequentialSampler:\n",
    "    * Instead of relying on the default random sampler, you can use SequentialSampler to ensure that the data is loaded in a sequential order, which might help in scenarios where the order of loading matters.\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "98abd35d-befe-40e6-ab4e-c678d48864ac",
   "metadata": {},
   "outputs": [],
   "source": [
    "trainloader = DataLoader(train_dataset, batch_size = CONFIG['batch_size'], shuffle = True, drop_last = True)\n",
    "\n",
    "testloader = DataLoader(test_dataset, batch_size = CONFIG['batch_size'], shuffle = False, drop_last = True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2c34038c-0762-4061-85ee-cafd44eddddf",
   "metadata": {},
   "source": [
    "# Train - Evaluation, Training, Architecture"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "fd3de3db-b023-41f9-b2d9-cf9f8c245a37",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Use `torch.no_grad()` here to disable gradient calculation. \n",
    "# It will reduce memory consumption as we don't need to compute gradients in inference.\n",
    "\n",
    "@torch.no_grad()\n",
    "def evaluate(model, testloader, criterion, device):\n",
    "    \n",
    "    model.eval()\n",
    "    \n",
    "    # Keep track of loss and other evaluation metrics\n",
    "    test_loss = 0.0\n",
    "\n",
    "    YPredict = []\n",
    "    YTrue = []\n",
    "    \n",
    "    # Loop through each batch on test set\n",
    "    for i, (images, labels) in enumerate(testloader, 1):\n",
    "        \n",
    "        images = images.to(device)\n",
    "        labels = labels.to(device)\n",
    "        \n",
    "        logits = model(images)\n",
    "\n",
    "        loss = criterion(logits, labels)\n",
    "        \n",
    "        test_loss += loss.detach().item()\n",
    "\n",
    "        predictions = torch.argmax(logits, dim = 1)\n",
    "        predictions = predictions.view(-1)\n",
    "        YTrue.extend(labels.tolist())\n",
    "        YPredict.extend(predictions.tolist())\n",
    "\n",
    "    # Compute Scores\n",
    "    accuracy = accuracy_score(YTrue, YPredict)\n",
    "    f1 = f1_score(YTrue, YPredict, average = \"macro\")\n",
    "    print(\"Classification report :\")\n",
    "    print(classification_report(YTrue, YPredict))\n",
    "\n",
    "    return (test_loss/i, accuracy, f1) ## avg loss and acc over all batches\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d9af0610-8e73-45cd-987c-9556427a514a",
   "metadata": {},
   "outputs": [],
   "source": [
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "9857e5e7-7f67-45c6-90f9-9a6586b650c8",
   "metadata": {},
   "outputs": [],
   "source": [
    "def train_model(model, trainloader, testloader, CONFIG):\n",
    "    \n",
    "    optimizer = AdamW(model.parameters(), lr = CONFIG[\"learning_rate\"], weight_decay = CONFIG[\"weight_decay\"])\n",
    "\n",
    "    # criterion = nn.CrossEntropyLoss(weight = class_weights_tensor)\n",
    "\n",
    "    criterion = nn.CrossEntropyLoss()\n",
    "\n",
    "\n",
    "    ## Measure runtime\n",
    "    t_start = time()\n",
    "    \n",
    "    ## Store training log\n",
    "    history = defaultdict(list)\n",
    "    \n",
    "    # We will train the model `num_epochs` times\n",
    "    for i in range(1, CONFIG[\"epochs\"]+1):\n",
    "\n",
    "        # Training\n",
    "        train_loss, train_acc, train_f1 = train_one_epoch(model = model, trainloader = trainloader, optimizer = optimizer, criterion = criterion, device = CONFIG[\"device\"])\n",
    "\n",
    "        # Model Checkpointing\n",
    "        if i%10==0:\n",
    "            torch.save(model.state_dict(), f'{CONFIG[\"model_save_folder\"]}epoch{i}_batch_size{CONFIG[\"batch_size\"]}.pth')\n",
    "\n",
    "        # Evaluate on test set\n",
    "        test_loss, test_acc, test_f1 = evaluate(model = model, testloader = testloader, criterion = criterion, device = CONFIG[\"device\"])\n",
    "\n",
    "        ## Store train/test loss, accuracy, f1 scores\n",
    "        history[\"train_loss\"].append(train_loss)\n",
    "        history[\"train_acc\"].append(train_acc)\n",
    "        history[\"train_f1\"].append(train_f1)\n",
    "        \n",
    "        history[\"test_loss\"].append(test_loss)\n",
    "        history[\"test_acc\"].append(test_acc)\n",
    "        history[\"test_f1\"].append(test_f1)\n",
    "\n",
    "        \n",
    "        \n",
    "        ## print out train/test loss, accuracy\n",
    "        print(f'Epoch: {i} | Runtime: {((time()-t_start)/60):.2f}[m] | train_loss: {train_loss:.3f} | test_loss: {test_loss:.3f} | train_acc: {train_acc:.3f} |  test_acc: {test_acc:.3f} | train_f1: {train_f1:.3f} | test_f1: {test_f1:.3f}')    \n",
    "    return history"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "44a71bce-186f-4c0c-9b32-308f9819afc4",
   "metadata": {},
   "outputs": [],
   "source": [
    "def train_one_epoch(model, trainloader, device, optimizer, criterion):\n",
    "    \"\"\"\n",
    "      Train 1 epoch on trainloader.\"\n",
    "    \"\"\"\n",
    "\n",
    "    model = model.train()\n",
    "    \n",
    "    # Keep track of loss and other evaluation metrics\n",
    "    train_loss = 0.0\n",
    "\n",
    "    YPredict = []\n",
    "    YTrue = []\n",
    "\n",
    "    ## Loop over all the batches\n",
    "    for i, (images, labels) in tqdm(enumerate(trainloader, 1), total=len(trainloader), desc=f\"Training on whole batch...\"):\n",
    "            \n",
    "        # Move images and labels to `device` (CPU or GPU)\n",
    "        images = images.to(device)\n",
    "        labels = labels.to(device)\n",
    "\n",
    "        logits = model(images)\n",
    "\n",
    "        # print(logits)\n",
    "        # print(labels)\n",
    "        # print(logits.dtype, labels.dtype)\n",
    "        # print(type(labels[0]))\n",
    "        # print(type(labels))\n",
    "        loss = criterion(logits, labels)\n",
    "        # loss = criterion(logits, labels.to(torch.long))\n",
    "        optimizer.zero_grad()\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "\n",
    "        # Compute loss and accuracy for this batch\n",
    "        train_loss += loss.detach().item()\n",
    "\n",
    "        predictions = torch.argmax(logits, dim = 1)\n",
    "        predictions = predictions.view(-1)\n",
    "        YTrue.extend(labels.tolist())\n",
    "        YPredict.extend(predictions.tolist())\n",
    "\n",
    "\n",
    "    # Compute Scores\n",
    "    accuracy = accuracy_score(YTrue, YPredict)\n",
    "    f1 = f1_score(YTrue, YPredict, average = \"macro\")\n",
    "\n",
    "    return (train_loss/i, accuracy, f1) ## avg loss and acc over all batches\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "id": "05d4166d-2d7d-4c79-a87c-9366953343ad",
   "metadata": {},
   "outputs": [],
   "source": [
    "class Baseline_Model(nn.Module):\n",
    "\n",
    "    def __init__(self):\n",
    "        super().__init__()\n",
    "        \n",
    "        # self.base_model_headless = base_layers\n",
    "        # self.base_model_headless.requires_grad = False\n",
    "        \n",
    "        self.fc1 = nn.Linear(CONFIG[\"feature_input_size\"], 128)\n",
    "        self.fc2 = nn.Linear(128, 32)\n",
    "        self.fc3 = nn.Linear(32, 5)\n",
    "        self.relu = nn.ReLU()\n",
    "        self.dropout = nn.Dropout(p = 0.1)\n",
    "\n",
    "\n",
    "    def forward(self, x):\n",
    "        # x = self.base_model_headless(x)\n",
    "        # # print(\"Output shape after Passing through Pretrained Model\")\n",
    "        # # print(x.shape)\n",
    "\n",
    "        # # Flatten\n",
    "        # x = x.view((x.shape[0], -1))\n",
    "\n",
    "        # FCNN\n",
    "        x = self.fc1(x)\n",
    "        x = self.relu(x)\n",
    "        x = self.dropout(x)\n",
    "        x = self.fc2(x)\n",
    "        x = self.relu(x)\n",
    "        x = self.dropout(x)\n",
    "        x = self.fc3(x)\n",
    "        # print(x.shape)\n",
    "\n",
    "        return x      "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "184b00e2-750e-40e8-9699-67daeaadde48",
   "metadata": {},
   "source": [
    "# Train the Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "id": "1a15c053-9780-47d1-882f-15fbbb06ac7c",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Initialize Model\n",
    "model = Baseline_Model().to(CONFIG[\"device\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "id": "6f956533-8f0b-4dc3-bbc6-03ac4975d84d",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------------------------------------------------\n",
      "        Layer (type)               Output Shape         Param #\n",
      "================================================================\n",
      "            Linear-1                   [-1, 64]          81,984\n",
      "              ReLU-2                   [-1, 64]               0\n",
      "           Dropout-3                   [-1, 64]               0\n",
      "            Linear-4                   [-1, 32]           2,080\n",
      "              ReLU-5                   [-1, 32]               0\n",
      "            Linear-6                    [-1, 5]             165\n",
      "================================================================\n",
      "Total params: 84,229\n",
      "Trainable params: 84,229\n",
      "Non-trainable params: 0\n",
      "----------------------------------------------------------------\n",
      "Input size (MB): 0.00\n",
      "Forward/backward pass size (MB): 0.00\n",
      "Params size (MB): 0.32\n",
      "Estimated Total Size (MB): 0.33\n",
      "----------------------------------------------------------------\n"
     ]
    }
   ],
   "source": [
    "summary(model, input_size = (CONFIG[\"feature_input_size\"],))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "id": "3b3268f2-b071-4f84-a307-a41e365ce456",
   "metadata": {},
   "outputs": [],
   "source": [
    "# ## compute accuracy score in a batch\n",
    "# def compute_accuracy(logits, labels, batch_size):\n",
    "#     ''' \n",
    "#       Obtain accuracy for a training batch \n",
    "#       logits: float Tensor, shape (batch_size, num_classes),  output from the model\n",
    "#       labels: Long Tensor, shape (batch_size, ), contains labels for the predictions\n",
    "#       batch_size: int, batch size\n",
    "\n",
    "#       Return accuracy for this batch, which should be a float number\n",
    "#     '''\n",
    "    \n",
    "#     # Write your code in this block ----------------\n",
    "    \n",
    "#     predictions = torch.argmax(logits, dim = 1)\n",
    "#     predictions = predictions.view(-1)\n",
    "    \n",
    "#     total_correct = (labels == predictions).sum().item()\n",
    "#     accuracy = total_correct/batch_size\n",
    "    \n",
    "#     return accuracy"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "cb41d874-840c-46ba-a732-b538cff27f27",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "id": "84b0e68d-1d4a-4885-92ee-29a9abf92b05",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training on whole batch...: 100%|██████████| 188/188 [00:07<00:00, 25.06it/s]\n",
      "/projectnb/cs640grp/students/avarshn/.conda/envs/ocean/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1471: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/projectnb/cs640grp/students/avarshn/.conda/envs/ocean/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1471: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/projectnb/cs640grp/students/avarshn/.conda/envs/ocean/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1471: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Classification report :\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.00      0.00      0.00        53\n",
      "           1       0.00      0.00      0.00        30\n",
      "           2       0.09      1.00      0.16        11\n",
      "           3       1.00      0.04      0.08        24\n",
      "           4       0.00      0.00      0.00        10\n",
      "\n",
      "    accuracy                           0.09       128\n",
      "   macro avg       0.22      0.21      0.05       128\n",
      "weighted avg       0.19      0.09      0.03       128\n",
      "\n",
      "Epoch: 1 | Runtime: 0.13[m] | train_loss: 1.617 | test_loss: 1.660 | train_acc: 0.189 |  test_acc: 0.094 | train_f1: 0.064 | test_f1: 0.048\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training on whole batch...: 100%|██████████| 188/188 [00:07<00:00, 25.12it/s]\n",
      "/projectnb/cs640grp/students/avarshn/.conda/envs/ocean/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1471: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/projectnb/cs640grp/students/avarshn/.conda/envs/ocean/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1471: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/projectnb/cs640grp/students/avarshn/.conda/envs/ocean/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1471: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Classification report :\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.00      0.00      0.00        53\n",
      "           1       0.00      0.00      0.00        30\n",
      "           2       0.09      1.00      0.16        11\n",
      "           3       1.00      0.04      0.08        24\n",
      "           4       0.00      0.00      0.00        10\n",
      "\n",
      "    accuracy                           0.09       128\n",
      "   macro avg       0.22      0.21      0.05       128\n",
      "weighted avg       0.19      0.09      0.03       128\n",
      "\n",
      "Epoch: 2 | Runtime: 0.25[m] | train_loss: 1.610 | test_loss: 1.645 | train_acc: 0.191 |  test_acc: 0.094 | train_f1: 0.070 | test_f1: 0.048\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training on whole batch...: 100%|██████████| 188/188 [00:07<00:00, 25.14it/s]\n",
      "/projectnb/cs640grp/students/avarshn/.conda/envs/ocean/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1471: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/projectnb/cs640grp/students/avarshn/.conda/envs/ocean/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1471: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/projectnb/cs640grp/students/avarshn/.conda/envs/ocean/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1471: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Classification report :\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.75      0.06      0.11        53\n",
      "           1       0.17      0.03      0.06        30\n",
      "           2       0.10      1.00      0.18        11\n",
      "           3       0.20      0.04      0.07        24\n",
      "           4       0.00      0.00      0.00        10\n",
      "\n",
      "    accuracy                           0.12       128\n",
      "   macro avg       0.24      0.23      0.08       128\n",
      "weighted avg       0.40      0.12      0.08       128\n",
      "\n",
      "Epoch: 3 | Runtime: 0.38[m] | train_loss: 1.603 | test_loss: 1.630 | train_acc: 0.201 |  test_acc: 0.125 | train_f1: 0.095 | test_f1: 0.081\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training on whole batch...: 100%|██████████| 188/188 [00:07<00:00, 25.07it/s]\n",
      "/projectnb/cs640grp/students/avarshn/.conda/envs/ocean/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1471: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/projectnb/cs640grp/students/avarshn/.conda/envs/ocean/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1471: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/projectnb/cs640grp/students/avarshn/.conda/envs/ocean/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1471: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Classification report :\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.69      0.21      0.32        53\n",
      "           1       0.16      0.10      0.12        30\n",
      "           2       0.13      1.00      0.23        11\n",
      "           3       0.20      0.08      0.12        24\n",
      "           4       0.00      0.00      0.00        10\n",
      "\n",
      "    accuracy                           0.21       128\n",
      "   macro avg       0.24      0.28      0.16       128\n",
      "weighted avg       0.37      0.21      0.20       128\n",
      "\n",
      "Epoch: 4 | Runtime: 0.51[m] | train_loss: 1.595 | test_loss: 1.615 | train_acc: 0.267 |  test_acc: 0.211 | train_f1: 0.193 | test_f1: 0.159\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training on whole batch...: 100%|██████████| 188/188 [00:07<00:00, 25.13it/s]\n",
      "/projectnb/cs640grp/students/avarshn/.conda/envs/ocean/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1471: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/projectnb/cs640grp/students/avarshn/.conda/envs/ocean/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1471: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/projectnb/cs640grp/students/avarshn/.conda/envs/ocean/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1471: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Classification report :\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.53      0.38      0.44        53\n",
      "           1       0.18      0.13      0.15        30\n",
      "           2       0.15      0.82      0.25        11\n",
      "           3       0.50      0.17      0.25        24\n",
      "           4       0.00      0.00      0.00        10\n",
      "\n",
      "    accuracy                           0.29       128\n",
      "   macro avg       0.27      0.30      0.22       128\n",
      "weighted avg       0.37      0.29      0.29       128\n",
      "\n",
      "Epoch: 5 | Runtime: 0.63[m] | train_loss: 1.587 | test_loss: 1.602 | train_acc: 0.292 |  test_acc: 0.289 | train_f1: 0.231 | test_f1: 0.219\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training on whole batch...: 100%|██████████| 188/188 [00:07<00:00, 25.10it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Classification report :\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.49      0.42      0.45        53\n",
      "           1       0.14      0.10      0.12        30\n",
      "           2       0.16      0.73      0.26        11\n",
      "           3       0.44      0.17      0.24        24\n",
      "           4       0.00      0.00      0.00        10\n",
      "\n",
      "    accuracy                           0.29       128\n",
      "   macro avg       0.25      0.28      0.21       128\n",
      "weighted avg       0.33      0.29      0.28       128\n",
      "\n",
      "Epoch: 6 | Runtime: 0.76[m] | train_loss: 1.578 | test_loss: 1.591 | train_acc: 0.341 |  test_acc: 0.289 | train_f1: 0.295 | test_f1: 0.214\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training on whole batch...: 100%|██████████| 188/188 [00:07<00:00, 25.07it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Classification report :\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.50      0.49      0.50        53\n",
      "           1       0.10      0.07      0.08        30\n",
      "           2       0.17      0.73      0.28        11\n",
      "           3       0.20      0.04      0.07        24\n",
      "           4       0.25      0.10      0.14        10\n",
      "\n",
      "    accuracy                           0.30       128\n",
      "   macro avg       0.24      0.29      0.21       128\n",
      "weighted avg       0.30      0.30      0.27       128\n",
      "\n",
      "Epoch: 7 | Runtime: 0.89[m] | train_loss: 1.570 | test_loss: 1.579 | train_acc: 0.352 |  test_acc: 0.297 | train_f1: 0.309 | test_f1: 0.213\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training on whole batch...: 100%|██████████| 188/188 [00:07<00:00, 25.12it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Classification report :\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.49      0.55      0.52        53\n",
      "           1       0.19      0.13      0.16        30\n",
      "           2       0.17      0.64      0.26        11\n",
      "           3       1.00      0.04      0.08        24\n",
      "           4       0.20      0.10      0.13        10\n",
      "\n",
      "    accuracy                           0.33       128\n",
      "   macro avg       0.41      0.29      0.23       128\n",
      "weighted avg       0.47      0.33      0.30       128\n",
      "\n",
      "Epoch: 8 | Runtime: 1.01[m] | train_loss: 1.561 | test_loss: 1.570 | train_acc: 0.383 |  test_acc: 0.328 | train_f1: 0.355 | test_f1: 0.230\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training on whole batch...: 100%|██████████| 188/188 [00:07<00:00, 25.11it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Classification report :\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.49      0.55      0.52        53\n",
      "           1       0.25      0.17      0.20        30\n",
      "           2       0.17      0.64      0.27        11\n",
      "           3       1.00      0.04      0.08        24\n",
      "           4       0.12      0.10      0.11        10\n",
      "\n",
      "    accuracy                           0.34       128\n",
      "   macro avg       0.41      0.30      0.24       128\n",
      "weighted avg       0.47      0.34      0.31       128\n",
      "\n",
      "Epoch: 9 | Runtime: 1.14[m] | train_loss: 1.553 | test_loss: 1.563 | train_acc: 0.395 |  test_acc: 0.336 | train_f1: 0.361 | test_f1: 0.237\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training on whole batch...: 100%|██████████| 188/188 [00:07<00:00, 25.10it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Classification report :\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.51      0.58      0.54        53\n",
      "           1       0.27      0.13      0.18        30\n",
      "           2       0.17      0.64      0.26        11\n",
      "           3       1.00      0.04      0.08        24\n",
      "           4       0.11      0.10      0.11        10\n",
      "\n",
      "    accuracy                           0.34       128\n",
      "   macro avg       0.41      0.30      0.23       128\n",
      "weighted avg       0.48      0.34      0.31       128\n",
      "\n",
      "Epoch: 10 | Runtime: 1.27[m] | train_loss: 1.543 | test_loss: 1.556 | train_acc: 0.412 |  test_acc: 0.344 | train_f1: 0.386 | test_f1: 0.234\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training on whole batch...: 100%|██████████| 188/188 [00:07<00:00, 25.15it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Classification report :\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.52      0.62      0.56        53\n",
      "           1       0.27      0.10      0.15        30\n",
      "           2       0.16      0.64      0.26        11\n",
      "           3       1.00      0.04      0.08        24\n",
      "           4       0.11      0.10      0.11        10\n",
      "\n",
      "    accuracy                           0.35       128\n",
      "   macro avg       0.41      0.30      0.23       128\n",
      "weighted avg       0.49      0.35      0.31       128\n",
      "\n",
      "Epoch: 11 | Runtime: 1.39[m] | train_loss: 1.533 | test_loss: 1.551 | train_acc: 0.423 |  test_acc: 0.352 | train_f1: 0.392 | test_f1: 0.231\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training on whole batch...: 100%|██████████| 188/188 [00:07<00:00, 25.15it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Classification report :\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.52      0.60      0.56        53\n",
      "           1       0.29      0.13      0.18        30\n",
      "           2       0.17      0.64      0.26        11\n",
      "           3       1.00      0.04      0.08        24\n",
      "           4       0.11      0.10      0.11        10\n",
      "\n",
      "    accuracy                           0.35       128\n",
      "   macro avg       0.42      0.30      0.24       128\n",
      "weighted avg       0.49      0.35      0.32       128\n",
      "\n",
      "Epoch: 12 | Runtime: 1.52[m] | train_loss: 1.524 | test_loss: 1.545 | train_acc: 0.433 |  test_acc: 0.352 | train_f1: 0.399 | test_f1: 0.238\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training on whole batch...: 100%|██████████| 188/188 [00:07<00:00, 25.12it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Classification report :\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.50      0.60      0.55        53\n",
      "           1       0.31      0.13      0.19        30\n",
      "           2       0.17      0.64      0.27        11\n",
      "           3       1.00      0.04      0.08        24\n",
      "           4       0.11      0.10      0.11        10\n",
      "\n",
      "    accuracy                           0.35       128\n",
      "   macro avg       0.42      0.30      0.24       128\n",
      "weighted avg       0.49      0.35      0.32       128\n",
      "\n",
      "Epoch: 13 | Runtime: 1.65[m] | train_loss: 1.516 | test_loss: 1.538 | train_acc: 0.438 |  test_acc: 0.352 | train_f1: 0.406 | test_f1: 0.238\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training on whole batch...: 100%|██████████| 188/188 [00:07<00:00, 25.15it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Classification report :\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.52      0.60      0.56        53\n",
      "           1       0.27      0.13      0.18        30\n",
      "           2       0.17      0.64      0.26        11\n",
      "           3       1.00      0.04      0.08        24\n",
      "           4       0.11      0.10      0.11        10\n",
      "\n",
      "    accuracy                           0.35       128\n",
      "   macro avg       0.41      0.30      0.24       128\n",
      "weighted avg       0.49      0.35      0.32       128\n",
      "\n",
      "Epoch: 14 | Runtime: 1.77[m] | train_loss: 1.506 | test_loss: 1.533 | train_acc: 0.451 |  test_acc: 0.352 | train_f1: 0.421 | test_f1: 0.238\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training on whole batch...: 100%|██████████| 188/188 [00:07<00:00, 25.14it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Classification report :\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.52      0.60      0.56        53\n",
      "           1       0.33      0.13      0.19        30\n",
      "           2       0.16      0.64      0.25        11\n",
      "           3       1.00      0.04      0.08        24\n",
      "           4       0.11      0.10      0.11        10\n",
      "\n",
      "    accuracy                           0.35       128\n",
      "   macro avg       0.42      0.30      0.24       128\n",
      "weighted avg       0.50      0.35      0.32       128\n",
      "\n",
      "Epoch: 15 | Runtime: 1.90[m] | train_loss: 1.497 | test_loss: 1.529 | train_acc: 0.451 |  test_acc: 0.352 | train_f1: 0.427 | test_f1: 0.237\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training on whole batch...: 100%|██████████| 188/188 [00:07<00:00, 25.12it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Classification report :\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.54      0.60      0.57        53\n",
      "           1       0.33      0.17      0.22        30\n",
      "           2       0.16      0.64      0.25        11\n",
      "           3       0.50      0.04      0.08        24\n",
      "           4       0.12      0.10      0.11        10\n",
      "\n",
      "    accuracy                           0.36       128\n",
      "   macro avg       0.33      0.31      0.25       128\n",
      "weighted avg       0.42      0.36      0.33       128\n",
      "\n",
      "Epoch: 16 | Runtime: 2.03[m] | train_loss: 1.487 | test_loss: 1.524 | train_acc: 0.467 |  test_acc: 0.359 | train_f1: 0.440 | test_f1: 0.247\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training on whole batch...: 100%|██████████| 188/188 [00:07<00:00, 25.17it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Classification report :\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.55      0.60      0.58        53\n",
      "           1       0.29      0.17      0.21        30\n",
      "           2       0.16      0.64      0.26        11\n",
      "           3       0.50      0.04      0.08        24\n",
      "           4       0.12      0.10      0.11        10\n",
      "\n",
      "    accuracy                           0.36       128\n",
      "   macro avg       0.33      0.31      0.25       128\n",
      "weighted avg       0.41      0.36      0.33       128\n",
      "\n",
      "Epoch: 17 | Runtime: 2.15[m] | train_loss: 1.475 | test_loss: 1.517 | train_acc: 0.475 |  test_acc: 0.359 | train_f1: 0.452 | test_f1: 0.247\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training on whole batch...: 100%|██████████| 188/188 [00:07<00:00, 25.12it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Classification report :\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.54      0.60      0.57        53\n",
      "           1       0.35      0.20      0.26        30\n",
      "           2       0.16      0.64      0.26        11\n",
      "           3       1.00      0.04      0.08        24\n",
      "           4       0.12      0.10      0.11        10\n",
      "\n",
      "    accuracy                           0.37       128\n",
      "   macro avg       0.44      0.32      0.26       128\n",
      "weighted avg       0.52      0.37      0.34       128\n",
      "\n",
      "Epoch: 18 | Runtime: 2.28[m] | train_loss: 1.467 | test_loss: 1.510 | train_acc: 0.485 |  test_acc: 0.367 | train_f1: 0.467 | test_f1: 0.255\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training on whole batch...: 100%|██████████| 188/188 [00:07<00:00, 25.11it/s]\n",
      "/projectnb/cs640grp/students/avarshn/.conda/envs/ocean/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1471: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/projectnb/cs640grp/students/avarshn/.conda/envs/ocean/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1471: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/projectnb/cs640grp/students/avarshn/.conda/envs/ocean/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1471: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Classification report :\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.55      0.60      0.58        53\n",
      "           1       0.38      0.20      0.26        30\n",
      "           2       0.15      0.64      0.25        11\n",
      "           3       0.00      0.00      0.00        24\n",
      "           4       0.12      0.10      0.11        10\n",
      "\n",
      "    accuracy                           0.36       128\n",
      "   macro avg       0.24      0.31      0.24       128\n",
      "weighted avg       0.34      0.36      0.33       128\n",
      "\n",
      "Epoch: 19 | Runtime: 2.40[m] | train_loss: 1.457 | test_loss: 1.508 | train_acc: 0.486 |  test_acc: 0.359 | train_f1: 0.468 | test_f1: 0.239\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training on whole batch...: 100%|██████████| 188/188 [00:07<00:00, 25.14it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Classification report :\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.53      0.60      0.57        53\n",
      "           1       0.43      0.20      0.27        30\n",
      "           2       0.16      0.64      0.25        11\n",
      "           3       0.00      0.00      0.00        24\n",
      "           4       0.22      0.20      0.21        10\n",
      "\n",
      "    accuracy                           0.37       128\n",
      "   macro avg       0.27      0.33      0.26       128\n",
      "weighted avg       0.35      0.37      0.34       128\n",
      "\n",
      "Epoch: 20 | Runtime: 2.53[m] | train_loss: 1.448 | test_loss: 1.500 | train_acc: 0.507 |  test_acc: 0.367 | train_f1: 0.488 | test_f1: 0.260\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/projectnb/cs640grp/students/avarshn/.conda/envs/ocean/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1471: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/projectnb/cs640grp/students/avarshn/.conda/envs/ocean/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1471: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n",
      "/projectnb/cs640grp/students/avarshn/.conda/envs/ocean/lib/python3.9/site-packages/sklearn/metrics/_classification.py:1471: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use `zero_division` parameter to control this behavior.\n",
      "  _warn_prf(average, modifier, msg_start, len(result))\n"
     ]
    }
   ],
   "source": [
    "# Train Model\n",
    "\n",
    "history_model = train_model(model, trainloader, testloader, CONFIG)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ad0b4e03-74a6-4f50-ab92-66ae5087fb36",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "fe441e3b-ca49-4169-9237-25c32dd4891b",
   "metadata": {},
   "source": [
    "# Plots"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "id": "4d6bb5c3-1575-43e9-94b0-35e272d47f45",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x14b410194520>"
      ]
     },
     "execution_count": 98,
     "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": [
    "## Plot train_loss, test_loss\n",
    "num_epochs = CONFIG['epochs']\n",
    "plt.plot(np.arange(num_epochs), history_model[\"train_loss\"], label='train loss')\n",
    "plt.plot(np.arange(num_epochs), history_model[\"test_loss\"], label='test loss')\n",
    "plt.xlabel('#epochs')\n",
    "plt.ylabel('loss')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "id": "7060a727-2c6c-4528-ba8b-ca08f4c4123e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x14b3dff1baf0>"
      ]
     },
     "execution_count": 99,
     "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": [
    "## Plot accuracy\n",
    "\n",
    "plt.plot(np.arange(num_epochs), history_model[\"train_acc\"], label='train acc')\n",
    "plt.plot(np.arange(num_epochs), history_model[\"test_acc\"], label='test acc')\n",
    "plt.xlabel('#epochs')\n",
    "plt.ylabel('accuracy')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "id": "82512b88-1775-4423-832e-7f1d865de437",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x14b3dfe8cb50>"
      ]
     },
     "execution_count": 100,
     "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": [
    "## Plot accuracy\n",
    "\n",
    "plt.plot(np.arange(num_epochs), history_model[\"train_f1\"], label='train f1')\n",
    "plt.plot(np.arange(num_epochs), history_model[\"test_f1\"], label='test f1')\n",
    "plt.xlabel('#epochs')\n",
    "plt.ylabel('f1')\n",
    "plt.legend()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "id": "b3f44965-b6e6-4932-a517-08239c27afb5",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Baseline_Model(\n",
       "  (fc1): Linear(in_features=1280, out_features=64, bias=True)\n",
       "  (fc2): Linear(in_features=64, out_features=32, bias=True)\n",
       "  (fc3): Linear(in_features=32, out_features=5, bias=True)\n",
       "  (relu): ReLU()\n",
       "  (dropout): Dropout(p=0.1, inplace=False)\n",
       ")"
      ]
     },
     "execution_count": 101,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5e5ddf34-6152-4ffd-9fdf-e0ff7d64c1c6",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 164,
   "id": "71b5f7d9-95af-4002-9b4e-f0a520d09dcd",
   "metadata": {},
   "outputs": [],
   "source": [
    "# torch.save(model.state_dict(), f'models/Baseline_model/state_dict_epoch{i}_batch_size{CONFIG[\"batch_size\"]}.pth')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d66a740e-4beb-46e0-8208-cf3d4da05782",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}