This is a repository for experiment tracking using tensorboard. Pytorch was used for creating models and pre-existing models were tested and compared on various datasets for image classification.
Additionally, a custom built ViT is also implemented which can be used for experiment tracking as well. Due to the required training time and hardware for training a ViT, after a custom built implementation, a pre-existing ViT is also used from torchvision.models
Also includes a deployment to huggingface using gradio