Releases: lucidrains/byol-pytorch
Releases · lucidrains/byol-pytorch
0.1.5
bump for release
0.1.4
make registration of hook lazy
0.1.3
bump package
0.1.2
Average loss rather than sum across batch
0.1.1
make sure to normalize channels by ImageNet statistics after aug
0.1.0
make sure NetWrapper class is pickle-able for distributed training. a…
0.0.5
fix issue with hidden representation coming from last layer
0.0.4
bump package
0.0.2
First release