This project implements a basic image classification model for the Fashion-MNIST dataset using a simple ResNet architecture and PyTorch Lightning.
Fashion-MNIST is a dataset of clothing images consisting of 70,000 images in 10 categories, commonly used for image classification tasks. This project demonstrates training a convolutional neural network (CNN) to classify fashion items from the Fashion-MNIST dataset using:
- Simple ResNet: A residual neural network architecture known for its effectiveness in image classification tasks.
- PyTorch Lightning: A framework built on top of PyTorch that simplifies the deep learning training process
This project requires the following dependencies:
- Python 3.x (> 3.6)
- PyTorch (https://pytorch.org/)
- torchvision (https://pypi.org/project/torchvision/)
- PyTorch Lightning (https://pytorch-lightning.readthedocs.io/en/1.3.3)
The training script allows you to customize various hyperparameters such as learning rate, batch size, and number of epochs. The evaluation script provides metrics like accuracy and loss on the test set.