- Data preprocessing and investigation
- Clean up the data labels.
- Make train, test and val splits for pretraining and transfer learning.
- Trimming pre-processing script
- Dataset definitions and tests
- Cleaning image data
- Add pad to square and resizing operations to homogenize the images
- add imagenet rgb norm
- add losses
- add network definitions
- add model definition
- first pretraining
- add confusion matrix
- add f-scores metric
- add focal loss
- add Flip Left Right Augmentation
- make inference script for test