CIFAR-10 Image Classification with PyTorch

Credit to https://pytorch.org/tutorials/beginner/blitz/cifar10_tutorial.html

Overview

This project implements a Convolutional Neural Network (CNN) using PyTorch to classify images from the CIFAR-10 dataset. The dataset consists of 60,000 32x32 color images in 10 classes, with 6,000 images per class. The model is trained to classify images into one of the following categories:

• Airplane
• Automobile
• Bird
• Cat
• Deer
• Dog
• Frog
• Horse
• Ship
• Truck

Dataset

The CIFAR-10 dataset is automatically downloaded and split into training and test sets.

Implementation Steps

1. Load and Normalize the CIFAR-10 Dataset

   • Use torchvision.datasets.CIFAR10 to load the dataset.
   • Apply transformations including normalization.
   • Use DataLoader to facilitate batch processing.

2. Visualize Sample Data

   • Display random training images using matplotlib.

3. Define the CNN Model

   • The model consists of:
     • Two convolutional layers (Conv2d)
     • Max pooling layers (MaxPool2d)
     • Fully connected layers (Linear)
     • Activation functions (ReLU)

4. Define the Loss Function and Optimizer

   • Use Cross-Entropy Loss (nn.CrossEntropyLoss) as the loss function.
   • Use Stochastic Gradient Descent (SGD) with momentum for optimization.

5. Train the Model

   • Train for multiple epochs.
   • Calculate loss at every 2000 mini-batches.

6. Test the Model

   • Predict labels for test images.
   • Compare predictions with ground truth.
   • Evaluate the overall model accuracy.

7. Analyze Model Performance

   • Compute accuracy for each class separately.

8. Save the Trained Model

   • Save model parameters using torch.save().

9. Train on GPU (if available)

   • Move model and data to CUDA device for acceleration.

Requirements

To run this project, install the following dependencies:

pip install torch torchvision matplotlib numpy

Future Improvements

• Increase model complexity (add more layers or use pretrained models like ResNet).
• Use data augmentation techniques to improve generalization.
• Fine-tune hyperparameters for better accuracy.
• Train for more epochs to improve performance.

---

This project provides a foundational approach to image classification using CNNs in PyTorch. 🚀