0.4.3 release
What's Changed
- [docs][fix] Update example to use offload_model by @anj-s in #806
- Switch default branch from master to main by @tmarkstrum in #807
- [FairScale] Remove refs to "cpu_offload" in code comments by @rohan-varma in #814
- [chore] Remove deprecated THCudaCheck by @anj-s in #818
- [feat] layer memory tracking by @QuentinDuval in #808
- [chore] Add log for the new experimental memory tracker feature. by @anj-s in #819
- [chore] Update the PyTorch version that we run CPU tests with by @anj-s in #809
- [chore] Update the PyTorch version that we run benchmarks with. by @anj-s in #823
- Extend auto shard capabilities to work around torch.fx edge cases. by @EugenHotaj in #817
- [fix] Update golden data for account for the speed regression by @anj-s in #825
- [chore] Fix main breakage temporarily by relaxing constraints by @anj-s in #828
- Use correct node names for param counting in auto_shard. by @EugenHotaj in #830
- [chore] Update requirements file to reflect latest config by @anj-s in #832
- [fix]: Fixes an issue with pre_backward hook registering by @min-xu-ai in #833
- [feature] Skip creating the CPU grad tensor when training by @anj-s in #821
- [test] improve a test's coverage by @min-xu-ai in #798
- [fix] Decouple
move_params_to_cpu
from themixed_precision
. by @anj-s in #822 - [fix] fix test on main by @min-xu-ai in #835
- [feature] Add the low level SSD APIs by @anj-s in #829
- [feat] [FSDP]: add experimental support to shared weights by @min-xu-ai in #836
- update nightly torch and test the flaky test by @min-xu-ai in #837
- [chore] Fix broken main due to updated github URL requirements by @anj-s in #838
- [chore] Update Sphinx version in docs requirements file by @vtantia in #841
- [feat] experimental MEVO layer by @min-xu-ai in #840
- [feat] Gossip/SlowMo by @blefaudeux in #378
- [feature]Add support for SSD offload with FSDP for eval workloads by @anj-s in #839
- [chore] 0.4.2 release by @anupambhatnagar in #846
- CI config changes by @anupambhatnagar in #847
- Setup pre-commit github action and apply pre-commit to all files by @anupambhatnagar in #849
- Allow sharded grad scaler to cpu offload with FSDP by @anupambhatnagar in #831
- Update changelog, removed meta.yml and requirements cleanup by @anupambhatnagar in #853
- [feature] Add a OffloadConfig object to specify offloading params to disk. by @anj-s in #855
- [POC] Testing Manual dispatch by @anupambhatnagar in #859
- [fix] [MEVO]: make mevo work with eval and optim_state checkpointing by @min-xu-ai in #851
- [chore] 0.4.3 release by @min-xu-ai in #860
New Contributors
- @rohan-varma made their first contribution in #814
- @EugenHotaj made their first contribution in #817
- @vtantia made their first contribution in #841
Full Changelog: v0.4.1...v0.4.3