notebooks/model_pre_training_Vit.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "#!pip install gdown\n",
    "#!pip install torch==2.1.0 torchvision==0.16.0 --index-url https://download.pytorch.org/whl/cu118"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Device: cuda:0\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<contextlib.ExitStack at 0x1f7f3179160>"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import sys\n",
    "sys.path.append('..')\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.backends.cudnn as cudnn\n",
    "import matplotlib.pyplot as plt\n",
    "from src.training import train_model, eval_func\n",
    "import os\n",
    "from torchvision import datasets, models, transforms\n",
    "from src.modelvis import visualize_models\n",
    "\n",
    "from src.models import ModelResnet152dTimm, VisionTransformerTimm\n",
    "from src.dataloader import MiniImageNetDataSet, createMiniImageNetDataLoaders, MiniImageNetDownloader\n",
    "import torchvision\n",
    "\n",
    "'''\n",
    "The line cudnn.benchmark = True is typically used in deep learning projects that utilize the CUDA Deep Neural Network (cuDNN) library.\n",
    "When cudnn.benchmark is set to True, it enables cuDNN to automatically find the best algorithm configuration for the specific \n",
    "input sizes and hardware being used. This can result in improved performance during training and inference.By enabling benchmarking, \n",
    "cuDNN will run a short benchmarking phase during the first iteration of the model to determine the optimal algorithm configuration. \n",
    "This configuration is then cached and used for subsequent iterations, leading to faster execution times.\n",
    "It's important to note that enabling benchmarking may introduce some overhead during the initial benchmarking phase, so it is typically \n",
    "recommended to use it when the input sizes are consistent throughout the training process.\n",
    "Overall, setting cudnn.benchmark to True can help optimize the performance of deep learning models that use cuDNN.\n",
    "'''\n",
    "torch.backends.cudnn.benchmark = True\n",
    "\n",
    "'''\n",
    "The line plt.ion() is a function call that activates interactive mode in matplotlib.\n",
    "When interactive mode is enabled, any plot that is created will be displayed immediately \n",
    "and can be updated dynamically. This means that you can modify the plot after it is displayed, \n",
    "such as changing the data or adding annotations, and the changes will be reflected in real-time.\n",
    "'''\n",
    "plt.ion()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.9.2\n",
      "1.8.2\n"
     ]
    }
   ],
   "source": [
    "print(torchvision.__version__)\n",
    "print(torch.__version__)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "device = torch.device(\"cuda:0\" if torch.cuda.is_available() else \"cpu\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2024-01-06 22:35:44,768 - INFO - MiniImageNet database already exists.\n"
     ]
    }
   ],
   "source": [
    "extractDir = '..\\\\data\\\\miniImageNet'\n",
    "url = 'https://drive.usercontent.google.com/download?id=107FTosYIeBn5QbynR46YG91nHcJ70whs&export=download&authuser=0&confirm=t&uuid=a50f1fb0-d0de-4900-905d-bab7437a3ec8&at=APZUnTXEKGryF0vQqVJF4LvHHCT1:1704557194939'\n",
    "downloader = MiniImageNetDownloader(url, extractDir)\n",
    "downloader.download()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_transforms = {\n",
    "    'train': transforms.Compose([\n",
    "        transforms.Resize(224),\n",
    "        transforms.ColorJitter(brightness=0.4, contrast=0.4, saturation=0.4),\n",
    "        transforms.RandomHorizontalFlip(),\n",
    "        transforms.GaussianBlur(3, sigma=(0.1, 2.0)),\n",
    "        transforms.ToTensor(),\n",
    "        transforms.Normalize((0.485, 0.456, 0.406), (0.229, 0.224, 0.225)),\n",
    "    ]),\n",
    "    'test': transforms.Compose([\n",
    "        transforms.Resize(224),\n",
    "        transforms.ToTensor(),\n",
    "        transforms.Normalize((0.485, 0.456, 0.406), (0.229, 0.224, 0.225)),\n",
    "    ]),\n",
    "    'val': transforms.Compose([\n",
    "        transforms.Resize(224),\n",
    "        transforms.ToTensor(),\n",
    "        transforms.Normalize((0.485, 0.456, 0.406), (0.229, 0.224, 0.225)),\n",
    "    ]),\n",
    "}\n",
    "dataDir = '..\\\\data\\\\miniImageNet\\\\train'\n",
    "dataloaders, class_names, dataset_sizes = createMiniImageNetDataLoaders(data_transforms, dataDir, split=0.8, batch_size=32)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Hyperparameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "LEARNING_RATE=0.001 # 0.0001\n",
    "WEIGHT_DECAY=0.0005 # 0.000001\n",
    "NUM_EPOCHS=10\n",
    "BATCH_SIZE=256\n",
    "MOMENTUM=0.9\n",
    "STEP_SIZE=7\n",
    "GAMMA=0.01"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "NUM_OF_CLASSES = 64\n",
    "net = VisionTransformerTimm(NUM_OF_CLASSES)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Learning Rate Control"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "for p in net.parameters():\n",
    "    p.requires_grad = True\n",
    "net.train()\n",
    "\n",
    "criterion = torch.nn.CrossEntropyLoss()\n",
    "optimizer = torch.optim.SGD(params= net.parameters(), lr=LEARNING_RATE, momentum=MOMENTUM, weight_decay=WEIGHT_DECAY)\n",
    "scheduler = torch.optim.lr_scheduler.StepLR(optimizer, step_size=STEP_SIZE, gamma=GAMMA)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Please wait patiently, it may take some minutes...\n",
      "Epoch 0/9\n",
      "----------\n",
      "train Loss: 0.0150 Acc: 0.8747\n",
      "val Loss: 0.0080 Acc: 0.9227\n",
      "\n",
      "Epoch 1/9\n",
      "----------\n",
      "train Loss: 0.0048 Acc: 0.9551\n",
      "val Loss: 0.0076 Acc: 0.9253\n",
      "\n",
      "Epoch 2/9\n",
      "----------\n",
      "train Loss: 0.0026 Acc: 0.9761\n",
      "val Loss: 0.0072 Acc: 0.9346\n",
      "\n",
      "Epoch 3/9\n",
      "----------\n",
      "train Loss: 0.0016 Acc: 0.9861\n",
      "val Loss: 0.0073 Acc: 0.9313\n",
      "\n",
      "Epoch 4/9\n",
      "----------\n",
      "train Loss: 0.0010 Acc: 0.9915\n",
      "val Loss: 0.0073 Acc: 0.9344\n",
      "\n",
      "Epoch 5/9\n",
      "----------\n",
      "train Loss: 0.0007 Acc: 0.9953\n",
      "val Loss: 0.0076 Acc: 0.9331\n",
      "\n",
      "Epoch 6/9\n",
      "----------\n",
      "train Loss: 0.0005 Acc: 0.9965\n",
      "val Loss: 0.0073 Acc: 0.9375\n",
      "\n",
      "Epoch 7/9\n",
      "----------\n",
      "train Loss: 0.0004 Acc: 0.9972\n",
      "val Loss: 0.0072 Acc: 0.9388\n",
      "\n",
      "Epoch 8/9\n",
      "----------\n",
      "train Loss: 0.0003 Acc: 0.9985\n",
      "val Loss: 0.0075 Acc: 0.9388\n",
      "\n",
      "Epoch 9/9\n",
      "----------\n",
      "train Loss: 0.0003 Acc: 0.9987\n",
      "val Loss: 0.0077 Acc: 0.9365\n",
      "\n",
      "Training complete in 99m 36s\n",
      "Best val Acc: 0.938802\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "print('Please wait patiently, it may take some minutes...')\n",
    "best_model = train_model(net, dataloaders, criterion, optimizer, scheduler, NUM_EPOCHS, dataset_sizes)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "save_path = '..\\\\data\\\\models\\\\best_model_VistionTransformerTimm_pc18_fixedTransformers.pth'\n",
    "torch.save(best_model.state_dict(), save_path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Accuracy on testing data: 0.938281\n"
     ]
    }
   ],
   "source": [
    "eval_acc = eval_func(best_model, dataloaders['test'])\n",
    "print('')\n",
    "print('Accuracy on testing data: %f' % eval_acc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "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": "3.9.16"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}