More object detection models and backbones

[akihironitta]

[0.5.0] - 2021-12-20

Added

• Added YOLO model (Add YOLO object detection model #552)
• Added SRGAN, SRImageLoggerCallback, TVTDataModule, SRCelebA, SRMNIST, SRSTL10 (Add SRGAN and datamodules for super resolution #466)
• Added nn.Module support for FasterRCNN backbone (Added nn.Module support for FasterRCNN backbone #661)
• Added RetinaNet with torchvision backbones (Add RetinaNet Object detection with Backbones #529)
• Added Python 3.9 support (Add Python 3.9 and drop PyTorch 1.6 #786)

Changed

• VAE now uses deterministic KL divergence during training, previously estimated KL divergence by random sampling (Replaced KL divergence estimation with deterministic KL #760)

Removed

• Removed PyTorch 1.6 support (Add Python 3.9 and drop PyTorch 1.6 #786)
• Removed Python 3.6 support (Drop Python 3.6 support #785)

Fixed

• Fixed doctest fails with ImportError: cannot import name 'Env' from 'gym' (Update experience_source.py #749 Bug Fix #751)
• Fixed MoCo v2 missing Cosine Annealing learning scheduler (fix: CosineAnnealingLR is missing #757)

Contributors

@abhayraw1 @akihironitta @chris-clem @hoangtnm @nmichlo @oke-aditya @Programmer-RD-AI @senarvi

If we forgot someone due to not matching commit email with GitHub account, let us know :]

---

This discussion was created from the release More object detection models and backbones.