This project demonstrates how to train a ResNet-34 model on the Fashion MNIST dataset using PyTorch. The goal is to classify 10 different classes of fashion images.
- ✅ Use of ResNet-34 architecture for image classification.
- ✅ Dataset: Fashion MNIST (grayscale 28x28 images with 10 fashion categories).
- ✅ Use of PyTorch, torchvision for model & data handling.
- ✅ Data augmentation during training.
- ✅ Model checkpoint saving.
- ✅ Inference code for test dataset.
- ✅ Output predictions can be saved to CSV.
- Dataset: Fashion MNIST
- You can download automatically using
torchvision.datasets.FashionMNIST, or download manually and store insidedata/directory.
Below is the loss function during training:
