Rishabh-9090/Galaxy_Morphology_Classification_using_CNNs_and_ViT

Update metrics in results

Add results metrics and confusion matrices as json files

Add save confusion matrices function to save all confusion matrices

Add the best VGG16 filepath

Update project structure

Add project structure

Add HUggingFace space link

Update README.md

Add ViT effectiveness on disturbed galaxies insight

Add results summary

Clean notebook

Update README with Hugging Face demo details

Fix the handling of model and transformation tuple while loading the model

Add clean UI features

Fix HF checkpoint paths (remove local checkpoints prefix)

Switch checkpoints to HF model repo paths

Remove checkpoints from repo; load weights externally

Allow checkpoints folder for Git LFS

Fix imports for HF Spaces

Fix imports

Fix PYTHONPATH for Hugging Face Spaces

Add working Gradio inference for CNN and ViT models

Add visualisation

TRain Custom CNN on balanced dataset using early stopping and evaluate on balanced and unbalanced test dataset

Train Custom CNN using early stopping and evaluation

Set persistent training as false and added pre fetch factor

Train VGG19 on balanced dataset using early stopping and evaluate on balanced and unbalanced test dataset

Train VGG19 using early stopping and evaluation

Train ViT on balanced dataset and evaluate on balanced and unbalanced test dataset

Train ViT on unbalanced dataset and evaluation

Add validation optional feature in train_vit function

Update dataloader function for ViT

Add train vit function

Train VGG16 on balanced dataset using early stopping and evaluate on balanced and unbalanced test dataset

Train VGG16 using early stopping and evaluation

Update early stopping class

Train ResNet18 on balanced dataset using early stopping and evaluate on balanced and unbalanced test dataset

Train ResNet18 using early stopping and evaluation

Train ResNet26 on balanced dataset and evaluate on balanced and unbalanced dataset

Train and evaluate ResNet26 using early stopping

Updated train function docstring comment

TRain ResNet50 on balanced dataset using early stopping and evaluate on balanced and unbalanced test dataset

Train ResNet50 using early stopping

Add epoch number in returned elements

Add early stopping in training pipeline

Add epoch tracking

Add early stopping feature

Clean import discrepancies

Train resnet18 on balanced dataset and evaluate using balanced and unbalanced test dataset

Train and evaluate resnet18