{ "cells": [ { "cell_type": "markdown", "id": "eb28146d", "metadata": { "id": "eb28146d" }, "source": [ "# PyTorch fine-tuning of ResNet50\n", "\n", "## Fine-tuning\n", "\n", "Fine-tuning takes a model that has already been trained on a large dataset (like ImageNet) and continues training it on a new, smaller, task-specific dataset, adapting its learned features rather than starting from randomly initialized weights. This is typically much faster and more data-efficient than training from scratch, especially when the new dataset is small, because the pretrained model already encodes generally useful visual features like edges, textures, and shapes.\n", "\n", "## Major image classification architecture advances (2000s-2020)\n", "\n", "| Year | Architecture | Params | Dataset | Top-1 accuracy | Key advance | Paper |\n", "|------|--------------|--------|---------|-----------------|-------------|-------|\n", "| 2006 | Deep Belief Networks | N/A | MNIST (60K train / 10K test images) | 98.75% | Layer-wise unsupervised pretraining made training deep networks feasible | [A Fast Learning Algorithm for Deep Belief Nets](https://www.cs.toronto.edu/~hinton/absps/fastnc.pdf) |\n", "| 2012 | AlexNet | ~60M | ImageNet ILSVRC-2012 (1.2M train images, 1000 classes) | 62.5% | Deep CNN + ReLU + dropout + GPU training won ImageNet by a huge margin, kicking off the deep learning boom | [ImageNet Classification with Deep Convolutional Neural Networks](https://papers.nips.cc/paper_files/paper/2012/file/c399862d3b9d6b76c8436e924a68c45b-Paper.pdf) |\n", "| 2014 | VGG | ~138M (VGG-16) | ImageNet ILSVRC-2012 (1.2M train images, 1000 classes) | 71.5% | Showed that stacking small 3x3 convolutions to build very deep, uniform networks improves accuracy | [Very Deep Convolutional Networks for Large-Scale Image Recognition](https://arxiv.org/pdf/1409.1556) |\n", "| 2014 | GoogLeNet (Inception) | ~6.8M | ImageNet ILSVRC-2014 (1.2M train images, 1000 classes) | 69.8% | Introduced inception modules with parallel multi-scale convolutions for efficient computation | [Going Deeper with Convolutions](https://arxiv.org/pdf/1409.4842) |\n", "| 2015 | ResNet | ~25.6M (ResNet-50) | ImageNet ILSVRC-2015 (1.2M train images, 1000 classes) | 76.2% | Introduced residual (skip) connections, enabling stable training of networks with 100+ layers | [Deep Residual Learning for Image Recognition](https://arxiv.org/pdf/1512.03385) |\n", "| 2017 | DenseNet | ~8M (DenseNet-121) | ImageNet (1.2M train images, 1000 classes) | 75.0% | Connected each layer to every other layer in a feed-forward fashion, improving feature reuse and gradient flow | [Densely Connected Convolutional Networks](https://arxiv.org/pdf/1608.06993) |\n", "| 2017 | MobileNet | ~4.2M | ImageNet (1.2M train images, 1000 classes) | 70.6% | Introduced depthwise separable convolutions for efficient inference on mobile/edge devices | [MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications](https://arxiv.org/pdf/1704.04861) |\n", "| 2019 | EfficientNet | ~5.3M (B0) | ImageNet (1.2M train images, 1000 classes) | 77.1% | Used compound scaling to balance network depth, width, and resolution for better accuracy/efficiency tradeoffs | [EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks](https://arxiv.org/pdf/1905.11946) |\n", "| 2020 | Vision Transformer (ViT) | ~86M (ViT-B/16) | ImageNet-21k pretrain + ImageNet-1k fine-tune (14M / 1.2M images) | 84.2% | Applied the transformer architecture directly to image patches, removing convolutions entirely | [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://arxiv.org/pdf/2010.11929) |\n", "\n", "## Notebook set-up\n", "\n", "### Dependencies\n", "\n", "**Note**: this notebook derived from the [`gperdrizet/CIFAR10`](https://github.com/gperdrizet/CIFAR10) repository and is meant to run in the development container environment specified in that repository. If you are attempting to run in somewhere else you may have to install some dependencies. Start with:\n", "\n", "```\n", "pip install image-classification-tools\n", "```\n", "\n", "Also install anything else that causes a 'module not found' error the same way.\n", "\n", "### Imports\n" ] }, { "cell_type": "code", "execution_count": 1, "id": "acc6f4eb", "metadata": { "colab": { "base_uri": "https://localhost:8080/", "height": 384 }, "executionInfo": { "elapsed": 13844, "status": "error", "timestamp": 1788494641377, "user": { "displayName": "George Perdrizet", "userId": "09442568899740404254" }, "user_tz": 240 }, "id": "acc6f4eb", "outputId": "84bc713a-c8ce-4f89-fb5a-10e11787d3b3" }, "outputs": [], "source": [ "# Standard library imports\n", "import pickle\n", "from pathlib import Path\n", "\n", "# Third party imports\n", "import matplotlib.pyplot as plt\n", "import numpy as np\n", "import torch\n", "import torch.nn as nn\n", "import torch.optim as optim\n", "from torchvision import datasets, transforms, models\n", "\n", "# Package imports\n", "from image_classification_tools.pytorch import DataPipeline\n", "import image_classification_tools.pytorch.evaluation as eval_utils\n", "import image_classification_tools.pytorch.plotting as plots\n", "import image_classification_tools.pytorch.training as training" ] }, { "cell_type": "code", "execution_count": 2, "id": "912e6f11", "metadata": { "executionInfo": { "elapsed": 185, "status": "aborted", "timestamp": 1788494641542, "user": { "displayName": "George Perdrizet", "userId": "09442568899740404254" }, "user_tz": 240 }, "id": "912e6f11" }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Using device: gpu\n", "Torch device is: cuda\n" ] } ], "source": [ "# Set random seeds for reproducibility\n", "torch.manual_seed(315)\n", "np.random.seed(315)\n", "\n", "# Device configuration\n", "device = 'gpu' if torch.cuda.is_available() else 'cpu'\n", "torch_device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n", "print(f'Using device: {device}')\n", "print(f'Torch device is: {torch_device}')\n", "\n", "# Paths\n", "models_dir = Path('../models/pytorch')\n", "results_dir = Path('../data/pytorch/performance_results')\n", "data_dir = Path('../data')\n", "\n", "# Make sure directories exist\n", "models_dir.mkdir(parents=True, exist_ok=True)\n", "results_dir.mkdir(parents=True, exist_ok=True)\n", "data_dir.mkdir(parents=True, exist_ok=True)\n", "\n", "# CIFAR-10 class names in class order\n", "class_names = [\n", " 'airplane', 'automobile', 'bird', 'cat', 'deer',\n", " 'dog', 'frog', 'horse', 'ship', 'truck'\n", "]" ] }, { "cell_type": "markdown", "id": "be9c1f7b", "metadata": { "id": "be9c1f7b" }, "source": [ "### Configuration" ] }, { "cell_type": "markdown", "id": "27433cdd", "metadata": { "id": "27433cdd" }, "source": [ "### Hyperparameters" ] }, { "cell_type": "code", "execution_count": 3, "id": "92f9ee0c", "metadata": { "executionInfo": { "elapsed": 197, "status": "aborted", "timestamp": 1788494641555, "user": { "displayName": "George Perdrizet", "userId": "09442568899740404254" }, "user_tz": 240 }, "id": "92f9ee0c" }, "outputs": [], "source": [ "validation_size = 10000\n", "batch_size = 64\n", "learning_rate = 1e-3\n", "epochs = 100 # Early stopping will halt if validation accuracy stops improving\n", "print_every = 5" ] }, { "cell_type": "markdown", "id": "1775dfbf", "metadata": { "id": "1775dfbf" }, "source": [ "## 1. Load and preprocess CIFAR-10 data\n", "\n", "### 1.1. Define data loading transforms" ] }, { "cell_type": "code", "execution_count": 4, "id": "6HUGRwDokvII", "metadata": { "executionInfo": { "elapsed": 196, "status": "aborted", "timestamp": 1788494641557, "user": { "displayName": "George Perdrizet", "userId": "09442568899740404254" }, "user_tz": 240 }, "id": "6HUGRwDokvII" }, "outputs": [], "source": [ "\n", "# Data preprocessing: convert to grayscale, tensor, and normalize\n", "transform = transforms.Compose([\n", " transforms.ToTensor(),\n", " transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5)),\n", "])" ] }, { "cell_type": "markdown", "id": "lYraCEmokw3U", "metadata": { "id": "lYraCEmokw3U" }, "source": [ "### 1.2. Download CIFAR10" ] }, { "cell_type": "code", "execution_count": 5, "id": "qqXy99ARk0Ya", "metadata": { "executionInfo": { "elapsed": 198, "status": "aborted", "timestamp": 1788494641559, "user": { "displayName": "George Perdrizet", "userId": "09442568899740404254" }, "user_tz": 240 }, "id": "qqXy99ARk0Ya" }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Files already downloaded and verified\n", "Files already downloaded and verified\n", "\n", "Training samples: 50000\n", "Test samples: 10000\n", "Image shape: torch.Size([3, 32, 32])\n", "Number of classes: 10\n" ] } ], "source": [ "# Use the fast BrainChip mirror for downloading CIFAR-10\n", "datasets.CIFAR10.url = 'https://data.brainchip.com/dataset-mirror/cifar10/cifar-10-python.tar.gz'\n", "\n", "# Load training and test datasets\n", "train_dataset = datasets.CIFAR10(\n", " root=data_dir,\n", " train=True,\n", " download=True,\n", " transform=transform\n", ")\n", "\n", "test_dataset = datasets.CIFAR10(\n", " root=data_dir,\n", " train=False,\n", " download=True,\n", " transform=transform\n", ")\n", "\n", "print(f'\\nTraining samples: {len(train_dataset)}')\n", "print(f'Test samples: {len(test_dataset)}')\n", "print(f'Image shape: {train_dataset[0][0].shape}')\n", "print(f'Number of classes: {len(class_names)}')" ] }, { "cell_type": "markdown", "id": "9oGEDJ3wkmEP", "metadata": { "id": "9oGEDJ3wkmEP" }, "source": [ "\n", "### 1.3. Visualize sample images" ] }, { "cell_type": "code", "execution_count": 6, "id": "10473ed2", "metadata": { "executionInfo": { "elapsed": 188, "status": "aborted", "timestamp": 1788494641560, "user": { "displayName": "George Perdrizet", "userId": "09442568899740404254" }, "user_tz": 240 }, "id": "10473ed2" }, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# Get a sample dataset for visualization\n", "sample_dataset = datasets.CIFAR10(root=data_dir, transform=transform)\n", "\n", "# Plot first 10 images from the training dataset\n", "fig, axes = plots.plot_sample_images(sample_dataset, class_names)\n", "plt.show()" ] }, { "cell_type": "markdown", "id": "efe7b605", "metadata": { "id": "efe7b605" }, "source": [ "### 1.4. Create PyTorch data loaders" ] }, { "cell_type": "code", "execution_count": 7, "id": "5c40812f", "metadata": { "executionInfo": { "elapsed": 191, "status": "aborted", "timestamp": 1788494641564, "user": { "displayName": "George Perdrizet", "userId": "09442568899740404254" }, "user_tz": 240 }, "id": "5c40812f" }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Loading train dataset...\n", "Loading test dataset...\n", "Creating splits...\n", "Preloading train to GPU:0...\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "Preloading train: 100%|██████████| 40000/40000 [00:17<00:00, 2307.26it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Preloading val to GPU:0...\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "Preloading val: 100%|██████████| 10000/10000 [00:04<00:00, 2438.50it/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "\n", "Batch sizes: {'train': 64, 'val': 64, 'test': 64}\n", "Total samples: {'train': 40000, 'val': 10000, 'test': 10000}\n", "Memory estimate: 0.69 GB\n" ] } ], "source": [ "# Create data pipeline with RGB preprocessing\n", "loaders = DataPipeline(\n", " data_source=datasets.CIFAR10,\n", " data_dir=data_dir,\n", " split='train/val/test',\n", " val_size=validation_size,\n", " batch_size=batch_size,\n", " train_transform=transform,\n", " eval_transform=transform,\n", " preload=device,\n", " seed=315\n", ").get_loaders()\n", "\n", "# Display pipeline summary\n", "print(f'\\nBatch sizes: {loaders.get_batch_sizes()}')\n", "print(f'Total samples: {loaders.total_samples()}')\n", "\n", "memory_estimate = loaders.memory_estimate()\n", "\n", "if memory_estimate is not None:\n", " print(f'Memory estimate: {memory_estimate:.2f} GB')\n", "\n", "else:\n", " print('Memory estimate: N/A')" ] }, { "cell_type": "markdown", "id": "8d86b258", "metadata": { "id": "8d86b258" }, "source": [ "## 2. ResNet50\n", "\n", "### 2.1. Load model" ] }, { "cell_type": "code", "execution_count": 8, "id": "894271d2", "metadata": { "executionInfo": { "elapsed": 191, "status": "aborted", "timestamp": 1788494641565, "user": { "displayName": "George Perdrizet", "userId": "09442568899740404254" }, "user_tz": 240 }, "id": "894271d2" }, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "Downloading: \"https://download.pytorch.org/models/resnet50-11ad3fa6.pth\" to /home/vscode/.cache/torch/hub/checkpoints/resnet50-11ad3fa6.pth\n", "100%|██████████| 97.8M/97.8M [00:01<00:00, 96.8MB/s]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "\n", "Model layers:\n", "\n", "Name: , Module Type: ResNet\n", "Name: conv1, Module Type: Conv2d\n", "Name: bn1, Module Type: BatchNorm2d\n", "Name: relu, Module Type: ReLU\n", "Name: maxpool, Module Type: MaxPool2d\n", "Name: layer1, Module Type: Sequential\n", "Name: layer1.0, Module Type: Bottleneck\n", "Name: layer1.0.conv1, Module Type: Conv2d\n", "Name: layer1.0.bn1, Module Type: BatchNorm2d\n", "Name: layer1.0.conv2, Module Type: Conv2d\n", "Name: layer1.0.bn2, Module Type: BatchNorm2d\n", "Name: layer1.0.conv3, Module Type: Conv2d\n", "Name: layer1.0.bn3, Module Type: BatchNorm2d\n", "Name: layer1.0.relu, Module Type: ReLU\n", "Name: layer1.0.downsample, Module Type: Sequential\n", "Name: layer1.0.downsample.0, Module Type: Conv2d\n", "Name: layer1.0.downsample.1, Module Type: BatchNorm2d\n", "Name: layer1.1, Module Type: Bottleneck\n", "Name: layer1.1.conv1, Module Type: Conv2d\n", "Name: layer1.1.bn1, Module Type: BatchNorm2d\n", "Name: layer1.1.conv2, Module Type: Conv2d\n", "Name: layer1.1.bn2, Module Type: BatchNorm2d\n", "Name: layer1.1.conv3, Module Type: Conv2d\n", "Name: layer1.1.bn3, Module Type: BatchNorm2d\n", "Name: layer1.1.relu, Module Type: ReLU\n", "Name: layer1.2, Module Type: Bottleneck\n", "Name: layer1.2.conv1, Module Type: Conv2d\n", "Name: layer1.2.bn1, Module Type: BatchNorm2d\n", "Name: layer1.2.conv2, Module Type: Conv2d\n", "Name: layer1.2.bn2, Module Type: BatchNorm2d\n", "Name: layer1.2.conv3, Module Type: Conv2d\n", "Name: layer1.2.bn3, Module Type: BatchNorm2d\n", "Name: layer1.2.relu, Module Type: ReLU\n", "Name: layer2, Module Type: Sequential\n", "Name: layer2.0, Module Type: Bottleneck\n", "Name: layer2.0.conv1, Module Type: Conv2d\n", "Name: layer2.0.bn1, Module Type: BatchNorm2d\n", "Name: layer2.0.conv2, Module Type: Conv2d\n", "Name: layer2.0.bn2, Module Type: BatchNorm2d\n", "Name: layer2.0.conv3, Module Type: Conv2d\n", "Name: layer2.0.bn3, Module Type: BatchNorm2d\n", "Name: layer2.0.relu, Module Type: ReLU\n", "Name: layer2.0.downsample, Module Type: Sequential\n", "Name: layer2.0.downsample.0, Module Type: Conv2d\n", "Name: layer2.0.downsample.1, Module Type: BatchNorm2d\n", "Name: layer2.1, Module Type: Bottleneck\n", "Name: layer2.1.conv1, Module Type: Conv2d\n", "Name: layer2.1.bn1, Module Type: BatchNorm2d\n", "Name: layer2.1.conv2, Module Type: Conv2d\n", "Name: layer2.1.bn2, Module Type: BatchNorm2d\n", "Name: layer2.1.conv3, Module Type: Conv2d\n", "Name: layer2.1.bn3, Module Type: BatchNorm2d\n", "Name: layer2.1.relu, Module Type: ReLU\n", "Name: layer2.2, Module Type: Bottleneck\n", "Name: layer2.2.conv1, Module Type: Conv2d\n", "Name: layer2.2.bn1, Module Type: BatchNorm2d\n", "Name: layer2.2.conv2, Module Type: Conv2d\n", "Name: layer2.2.bn2, Module Type: BatchNorm2d\n", "Name: layer2.2.conv3, Module Type: Conv2d\n", "Name: layer2.2.bn3, Module Type: BatchNorm2d\n", "Name: layer2.2.relu, Module Type: ReLU\n", "Name: layer2.3, Module Type: Bottleneck\n", "Name: layer2.3.conv1, Module Type: Conv2d\n", "Name: layer2.3.bn1, Module Type: BatchNorm2d\n", "Name: layer2.3.conv2, Module Type: Conv2d\n", "Name: layer2.3.bn2, Module Type: BatchNorm2d\n", "Name: layer2.3.conv3, Module Type: Conv2d\n", "Name: layer2.3.bn3, Module Type: BatchNorm2d\n", "Name: layer2.3.relu, Module Type: ReLU\n", "Name: layer3, Module Type: Sequential\n", "Name: layer3.0, Module Type: Bottleneck\n", "Name: layer3.0.conv1, Module Type: Conv2d\n", "Name: layer3.0.bn1, Module Type: BatchNorm2d\n", "Name: layer3.0.conv2, Module Type: Conv2d\n", "Name: layer3.0.bn2, Module Type: BatchNorm2d\n", "Name: layer3.0.conv3, Module Type: Conv2d\n", "Name: layer3.0.bn3, Module Type: BatchNorm2d\n", "Name: layer3.0.relu, Module Type: ReLU\n", "Name: layer3.0.downsample, Module Type: Sequential\n", "Name: layer3.0.downsample.0, Module Type: Conv2d\n", "Name: layer3.0.downsample.1, Module Type: BatchNorm2d\n", "Name: layer3.1, Module Type: Bottleneck\n", "Name: layer3.1.conv1, Module Type: Conv2d\n", "Name: layer3.1.bn1, Module Type: BatchNorm2d\n", "Name: layer3.1.conv2, Module Type: Conv2d\n", "Name: layer3.1.bn2, Module Type: BatchNorm2d\n", "Name: layer3.1.conv3, Module Type: Conv2d\n", "Name: layer3.1.bn3, Module Type: BatchNorm2d\n", "Name: layer3.1.relu, Module Type: ReLU\n", "Name: layer3.2, Module Type: Bottleneck\n", "Name: layer3.2.conv1, Module Type: Conv2d\n", "Name: layer3.2.bn1, Module Type: BatchNorm2d\n", "Name: layer3.2.conv2, Module Type: Conv2d\n", "Name: layer3.2.bn2, Module Type: BatchNorm2d\n", "Name: layer3.2.conv3, Module Type: Conv2d\n", "Name: layer3.2.bn3, Module Type: BatchNorm2d\n", "Name: layer3.2.relu, Module Type: ReLU\n", "Name: layer3.3, Module Type: Bottleneck\n", "Name: layer3.3.conv1, Module Type: Conv2d\n", "Name: layer3.3.bn1, Module Type: BatchNorm2d\n", "Name: layer3.3.conv2, Module Type: Conv2d\n", "Name: layer3.3.bn2, Module Type: BatchNorm2d\n", "Name: layer3.3.conv3, Module Type: Conv2d\n", "Name: layer3.3.bn3, Module Type: BatchNorm2d\n", "Name: layer3.3.relu, Module Type: ReLU\n", "Name: layer3.4, Module Type: Bottleneck\n", "Name: layer3.4.conv1, Module Type: Conv2d\n", "Name: layer3.4.bn1, Module Type: BatchNorm2d\n", "Name: layer3.4.conv2, Module Type: Conv2d\n", "Name: layer3.4.bn2, Module Type: BatchNorm2d\n", "Name: layer3.4.conv3, Module Type: Conv2d\n", "Name: layer3.4.bn3, Module Type: BatchNorm2d\n", "Name: layer3.4.relu, Module Type: ReLU\n", "Name: layer3.5, Module Type: Bottleneck\n", "Name: layer3.5.conv1, Module Type: Conv2d\n", "Name: layer3.5.bn1, Module Type: BatchNorm2d\n", "Name: layer3.5.conv2, Module Type: Conv2d\n", "Name: layer3.5.bn2, Module Type: BatchNorm2d\n", "Name: layer3.5.conv3, Module Type: Conv2d\n", "Name: layer3.5.bn3, Module Type: BatchNorm2d\n", "Name: layer3.5.relu, Module Type: ReLU\n", "Name: layer4, Module Type: Sequential\n", "Name: layer4.0, Module Type: Bottleneck\n", "Name: layer4.0.conv1, Module Type: Conv2d\n", "Name: layer4.0.bn1, Module Type: BatchNorm2d\n", "Name: layer4.0.conv2, Module Type: Conv2d\n", "Name: layer4.0.bn2, Module Type: BatchNorm2d\n", "Name: layer4.0.conv3, Module Type: Conv2d\n", "Name: layer4.0.bn3, Module Type: BatchNorm2d\n", "Name: layer4.0.relu, Module Type: ReLU\n", "Name: layer4.0.downsample, Module Type: Sequential\n", "Name: layer4.0.downsample.0, Module Type: Conv2d\n", "Name: layer4.0.downsample.1, Module Type: BatchNorm2d\n", "Name: layer4.1, Module Type: Bottleneck\n", "Name: layer4.1.conv1, Module Type: Conv2d\n", "Name: layer4.1.bn1, Module Type: BatchNorm2d\n", "Name: layer4.1.conv2, Module Type: Conv2d\n", "Name: layer4.1.bn2, Module Type: BatchNorm2d\n", "Name: layer4.1.conv3, Module Type: Conv2d\n", "Name: layer4.1.bn3, Module Type: BatchNorm2d\n", "Name: layer4.1.relu, Module Type: ReLU\n", "Name: layer4.2, Module Type: Bottleneck\n", "Name: layer4.2.conv1, Module Type: Conv2d\n", "Name: layer4.2.bn1, Module Type: BatchNorm2d\n", "Name: layer4.2.conv2, Module Type: Conv2d\n", "Name: layer4.2.bn2, Module Type: BatchNorm2d\n", "Name: layer4.2.conv3, Module Type: Conv2d\n", "Name: layer4.2.bn3, Module Type: BatchNorm2d\n", "Name: layer4.2.relu, Module Type: ReLU\n", "Name: avgpool, Module Type: AdaptiveAvgPool2d\n", "Name: fc, Module Type: Linear\n" ] } ], "source": [ "model = models.resnet50(weights=models.ResNet50_Weights.IMAGENET1K_V2)\n", "\n", "print('\\nModel layers:\\n')\n", "\n", "for name, module in model.named_modules():\n", " print(f'Name: {name}, Module Type: {module.__class__.__name__}')" ] }, { "cell_type": "markdown", "id": "391e180d", "metadata": { "id": "391e180d" }, "source": [ "### 2.2. Replace input\n", "\n", "The the first four layers in ResNet50 are a 7×7 convolution with a stride of 2 + max pooling with stride 2. This aggressively downsamples by 4x and was designed for 224×224 images. For 32×32 CIFAR images we can replace it with a 3×3, stride 1 convolution and remove the maxpool." ] }, { "cell_type": "code", "execution_count": 9, "id": "a33d9ea7", "metadata": { "executionInfo": { "elapsed": 193, "status": "aborted", "timestamp": 1788494641568, "user": { "displayName": "George Perdrizet", "userId": "09442568899740404254" }, "user_tz": 240 }, "id": "a33d9ea7" }, "outputs": [], "source": [ "# Replace the stem (conv1 + maxpool) to work with 32x32 CIFAR images\n", "# Original stem: 7x7 conv stride 2 + 3x3 maxpool stride 2 = 4x downsampling\n", "# New stem: 3x3 conv stride 1, no maxpool = preserves spatial resolution\n", "model.conv1 = nn.Conv2d(3, 64, kernel_size=3, stride=1, padding=1)\n", "model.maxpool = nn.Identity()" ] }, { "cell_type": "markdown", "id": "4826e4ff", "metadata": { "id": "4826e4ff" }, "source": [ "### 2.3. Replace output layer" ] }, { "cell_type": "code", "execution_count": 10, "id": "68f5c80e", "metadata": { "executionInfo": { "elapsed": 194, "status": "aborted", "timestamp": 1788494641570, "user": { "displayName": "George Perdrizet", "userId": "09442568899740404254" }, "user_tz": 240 }, "id": "68f5c80e" }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Original final layer input features: 2048\n", "Original final layer output: 1000 classes\n" ] } ], "source": [ "# Check the number of input/output features to the final layer\n", "num_inputs = model.fc.in_features\n", "num_outputs = model.fc.out_features\n", "print(f'Original final layer input features: {num_inputs}')\n", "print(f'Original final layer output: {num_outputs} classes')" ] }, { "cell_type": "code", "execution_count": 11, "id": "73e21e03", "metadata": { "executionInfo": { "elapsed": 36, "status": "aborted", "timestamp": 1788494641572, "user": { "displayName": "George Perdrizet", "userId": "09442568899740404254" }, "user_tz": 240 }, "id": "73e21e03" }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "New final layer: Dropout(0.5) -> Linear(2048, 10)\n" ] } ], "source": [ "# Replace the final layer with dropout + linear\n", "# Dropout helps prevent overfitting during fine-tuning\n", "model.fc = nn.Sequential(\n", " nn.Dropout(0.5),\n", " nn.Linear(num_inputs, len(class_names))\n", ")\n", "\n", "new_num_outputs = len(class_names)\n", "print(f'New final layer: Dropout(0.5) -> Linear({num_inputs}, {new_num_outputs})')" ] }, { "cell_type": "markdown", "id": "c11096c6", "metadata": { "id": "c11096c6" }, "source": [ "### 2.4. Freeze layers\n", "\n", "We freeze everything except:\n", "- The new stem (conv1) - since it's a new untrained layer\n", "- Layer 4 - the last set of 3 blocks of 3 convolutional layers\n", "- The fully connected output layer (fc)" ] }, { "cell_type": "code", "execution_count": 12, "id": "c6793089", "metadata": { "executionInfo": { "elapsed": 36, "status": "aborted", "timestamp": 1788494641573, "user": { "displayName": "George Perdrizet", "userId": "09442568899740404254" }, "user_tz": 240 }, "id": "c6793089" }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Total parameters: 23,520,906\n", "Trainable parameters: 16,695,050\n" ] } ], "source": [ "for name, param in model.named_parameters():\n", "\n", " # Freeze everything except conv1, layer4, and fc\n", " if 'conv1' not in name and 'layer4' not in name and 'fc' not in name:\n", " param.requires_grad = False\n", "\n", "total_params = sum(p.numel() for p in model.parameters())\n", "trainable_params = sum(p.numel() for p in model.parameters() if p.requires_grad)\n", "\n", "print(f'Total parameters: {total_params:,}')\n", "print(f'Trainable parameters: {trainable_params:,}')" ] }, { "cell_type": "markdown", "id": "082eb310", "metadata": { "id": "082eb310" }, "source": [ "## 3. Fine-tuning" ] }, { "cell_type": "code", "execution_count": 13, "id": "026dab17", "metadata": { "executionInfo": { "elapsed": 39, "status": "aborted", "timestamp": 1788494641577, "user": { "displayName": "George Perdrizet", "userId": "09442568899740404254" }, "user_tz": 240 }, "id": "026dab17" }, "outputs": [], "source": [ "# Define criterion and optimizer for training\n", "# Weight decay (L2 regularization) helps prevent overfitting\n", "criterion = nn.CrossEntropyLoss()\n", "optimizer = optim.Adam(model.parameters(), lr=learning_rate, weight_decay=1e-4)" ] }, { "cell_type": "code", "execution_count": 14, "id": "b96b3b19", "metadata": { "executionInfo": { "elapsed": 39, "status": "aborted", "timestamp": 1788494641578, "user": { "displayName": "George Perdrizet", "userId": "09442568899740404254" }, "user_tz": 240 }, "id": "b96b3b19" }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Epoch 1/100 - loss: 0.7282 - acc: 74.87% - val_loss: 0.3823 - val_acc: 87.15%\n", "Epoch 5/100 - loss: 0.1256 - acc: 95.75% - val_loss: 0.3176 - val_acc: 91.04%\n", "\n", "Early stopping at epoch 9\n", "Best val_loss: 0.2583 at epoch 6\n", "Restored best model weights\n", "\n", "CPU times: user 11min 14s, sys: 3.01 s, total: 11min 17s\n", "Wall time: 11min 15s\n" ] } ], "source": [ "%%time\n", "\n", "# Do the run, moving the model to the selected device\n", "# Early stopping will halt training when validation accuracy stops improving\n", "history = training.train_model(\n", " model=model.to(torch_device),\n", " train_loader=loaders.train,\n", " val_loader=loaders.val,\n", " criterion=criterion,\n", " optimizer=optimizer,\n", " lazy_loading=False,\n", " device=device,\n", " epochs=epochs,\n", " print_every=print_every,\n", " enable_early_stopping=True,\n", " early_stopping_patience=3\n", ")\n", "\n", "print()" ] }, { "cell_type": "code", "execution_count": 15, "id": "da355d78", "metadata": { "executionInfo": { "elapsed": 40, "status": "aborted", "timestamp": 1788494641580, "user": { "displayName": "George Perdrizet", "userId": "09442568899740404254" }, "user_tz": 240 }, "id": "da355d78" }, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "fig, axes = plots.plot_learning_curves(history)\n", "plt.show()" ] }, { "cell_type": "markdown", "id": "7da34565", "metadata": { "id": "7da34565" }, "source": [ "## 3. Evaluate model on test set\n", "\n", "### 3.1. Calculate test accuracy" ] }, { "cell_type": "code", "execution_count": 16, "id": "0323b059", "metadata": { "executionInfo": { "elapsed": 41, "status": "aborted", "timestamp": 1788494641582, "user": { "displayName": "George Perdrizet", "userId": "09442568899740404254" }, "user_tz": 240 }, "id": "0323b059" }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Test accuracy: 91.11%\n" ] } ], "source": [ "test_accuracy, predictions, true_labels = eval_utils.evaluate_model(model, loaders.test)\n", "print(f'Test accuracy: {test_accuracy:.2f}%')" ] }, { "cell_type": "markdown", "id": "bf993d13", "metadata": { "id": "bf993d13" }, "source": [ "### 3.2. Per-class accuracy" ] }, { "cell_type": "code", "execution_count": 17, "id": "af3fea3a", "metadata": { "executionInfo": { "elapsed": 43, "status": "aborted", "timestamp": 1788494641584, "user": { "displayName": "George Perdrizet", "userId": "09442568899740404254" }, "user_tz": 240 }, "id": "af3fea3a" }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Per-class accuracy:\n", "------------------------------\n", "airplane : 90.20%\n", "automobile : 95.30%\n", "bird : 91.00%\n", "cat : 78.10%\n", "deer : 93.40%\n", "dog : 87.10%\n", "frog : 93.70%\n", "horse : 95.00%\n", "ship : 91.80%\n", "truck : 95.50%\n" ] } ], "source": [ "# Calculate per-class accuracy\n", "class_correct = {name: 0 for name in class_names}\n", "class_total = {name: 0 for name in class_names}\n", "\n", "for pred, true in zip(predictions, true_labels):\n", "\n", " class_name = class_names[true]\n", " class_total[class_name] += 1\n", "\n", " if pred == true:\n", " class_correct[class_name] += 1\n", "\n", "print('Per-class accuracy:')\n", "print('-' * 30)\n", "\n", "for name in class_names:\n", " acc = 100 * class_correct[name] / class_total[name]\n", " print(f'{name:12s}: {acc:.2f}%')" ] }, { "cell_type": "markdown", "id": "19c6aa2d", "metadata": { "id": "19c6aa2d" }, "source": [ "### 3.3. Confusion matrix" ] }, { "cell_type": "code", "execution_count": 18, "id": "0fa8603b", "metadata": { "executionInfo": { "elapsed": 10, "status": "aborted", "timestamp": 1788494641586, "user": { "displayName": "George Perdrizet", "userId": "09442568899740404254" }, "user_tz": 240 }, "id": "0fa8603b" }, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "fig, ax = plots.plot_confusion_matrix(true_labels, predictions, class_names)\n", "plt.show()" ] }, { "cell_type": "markdown", "id": "51a080ec", "metadata": { "id": "51a080ec" }, "source": [ "### 3.4. Predicted class probability distributions" ] }, { "cell_type": "code", "execution_count": 19, "id": "11b5023e", "metadata": { "executionInfo": { "elapsed": 3, "status": "aborted", "timestamp": 1788494641588, "user": { "displayName": "George Perdrizet", "userId": "09442568899740404254" }, "user_tz": 240 }, "id": "11b5023e" }, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# Get predicted probabilities for all test samples\n", "model.eval()\n", "all_probs = []\n", "\n", "with torch.no_grad():\n", " for images, _ in loaders.test:\n", " outputs = model(images.to(torch_device))\n", " probs = torch.softmax(outputs, dim=1)\n", " all_probs.append(probs.cpu().numpy())\n", "\n", "all_probs = np.concatenate(all_probs, axis=0)\n", "\n", "# Plot probability distributions\n", "fig, axes = plots.plot_class_probability_distributions(all_probs, class_names)\n", "plt.show()" ] }, { "cell_type": "markdown", "id": "3109dd90", "metadata": { "id": "3109dd90" }, "source": [ "### 3.5. Evaluation curves" ] }, { "cell_type": "code", "execution_count": 20, "id": "8420d418", "metadata": { "executionInfo": { "elapsed": 14214, "status": "aborted", "timestamp": 1788494641589, "user": { "displayName": "George Perdrizet", "userId": "09442568899740404254" }, "user_tz": 240 }, "id": "8420d418" }, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "fig, (ax1, ax2) = plots.plot_evaluation_curves(true_labels, all_probs, class_names)\n", "plt.show()" ] }, { "cell_type": "markdown", "id": "a44452fa", "metadata": { "id": "a44452fa" }, "source": [ "## 4. Save model" ] }, { "cell_type": "code", "execution_count": 21, "id": "d4cc7d40", "metadata": { "executionInfo": { "elapsed": 14215, "status": "aborted", "timestamp": 1788494641591, "user": { "displayName": "George Perdrizet", "userId": "09442568899740404254" }, "user_tz": 240 }, "id": "d4cc7d40" }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Model saved to: ../models/pytorch/resnet50.pth\n", "Test accuracy: 91.11%\n" ] } ], "source": [ "# Save trained model\n", "model_path = models_dir / 'resnet50.pth'\n", "torch.save(model, model_path)\n", "\n", "print(f'Model saved to: {model_path}')\n", "print(f'Test accuracy: {test_accuracy:.2f}%')" ] }, { "cell_type": "markdown", "id": "86b87786", "metadata": { "id": "86b87786" }, "source": [ "## 5. Save test results for comparison" ] }, { "cell_type": "code", "execution_count": 22, "id": "6155613c", "metadata": { "executionInfo": { "elapsed": 14216, "status": "aborted", "timestamp": 1788494641592, "user": { "displayName": "George Perdrizet", "userId": "09442568899740404254" }, "user_tz": 240 }, "id": "6155613c" }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Test results saved to: ../data/pytorch/performance_results/resnet50_results.pkl\n", " - Test accuracy: 91.11%\n", " - Trainable parameters: 16,695,050\n", " - Total parameters: 23,520,906\n" ] } ], "source": [ "# Save results path\n", "results_path = results_dir / 'resnet50_results.pkl'\n", "\n", "# Count model parameters\n", "total_params = sum(p.numel() for p in model.parameters())\n", "trainable_params = sum(p.numel() for p in model.parameters() if p.requires_grad)\n", "\n", "# Create results dictionary\n", "results_dict = {\n", " 'true_labels': true_labels,\n", " 'predictions': predictions,\n", " 'all_probs': all_probs,\n", " 'test_accuracy': test_accuracy,\n", " 'total_params': total_params,\n", " 'trainable_params': trainable_params\n", "}\n", "\n", "# Save results\n", "with open(results_path, 'wb') as f:\n", " pickle.dump(results_dict, f)\n", "\n", "print(f'Test results saved to: {results_path}')\n", "\n", "print(f' - Test accuracy: {test_accuracy:.2f}%')\n", "print(f' - Trainable parameters: {trainable_params:,}')\n", "print(f' - Total parameters: {total_params:,}')\n" ] } ], "metadata": { "accelerator": "GPU", "colab": { "gpuType": "T4", "provenance": [ { "file_id": "https://github.com/gperdrizet/pytorch-demo/blob/main/notebooks/lesson-33-demo.ipynb", "timestamp": 1788485943007 } ] }, "kernelspec": { "display_name": "Python 3", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.10.12" } }, "nbformat": 4, "nbformat_minor": 5 }