ResNet10 Run Report
Summarized Findings
We introduced:
- EMA moving weight averages
- CutUp and MixUp, with an alpha of 0.1
Possible Improvements.
- we can add a fine tuning step where we reduce the augmentation complexity (no mixup/cutmix)
Notebook
Please see the companion notebook. It contains:
- Training Confusion Matrix
- Validation Confusion Matrix
- Training Classification Report
- Validation Classification Report
- Training GradCam
- Validation GradCam