Final project, Master in Artificial Intelligence, UPC BarcelonaTech
Device switching recommendations
Bug fix: channel augmentation
Fix bug when loading targets
Initialize with CleanFeatures instance
Resize input can be PIL Image or Tensor
Normalize range per channel after resize
Prevent division by zero
No channel augmentation by default
Resize readme
Fix tensor W, H confusion
Match number of channels to output requirements
Improve Resize constructor arguments
Check for Tensor input
Support datasets without targets
Add compute_feature_statistics documentation
Correct model-specific constructor arguments
Example use case
Fix type hints
Update and improve README