Super-gradients

This GitHub Release was done automatically by CircleCI

Features:
- KD trainer and resnet50 recipe (81.92 % accuracy)
- Repeated augmentation sampler
- Cooldown epochs
- Beit architecture
- Lamb Optimizer
- Passing torch data loaders directly to SgModel.

Refactoring:
- Checkpoint and architecture params decoupled.

- STDC architectures binary segmentation support bug fix.

torchmetrics version requirement set to 0.7.3 to avoid metric names depreciation bug.

This GitHub Release was done automatically by CircleCI

This GitHub Release was done automatically by CircleCI