Pretrain, finetune ANY AI model of ANY size on 1 or 10,000+ GPUs with zero code changes.
debug failing tests for Fabric with `ddp_fork` on PT 2.8 -> revert #21057 (#21092)
* debug failing tests for Fabric with `ddp_fork` on PT 2.8 * Revert "let `_get_default_process_group_backend_for_device` support more hardware platforms (#21057)" This reverts commit 119a640e43ee676d8491609f739a31b69857f4fe. --------- Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com> (cherry picked from commit 9ca360b1c6ba2dedbd77e9a559512b2f5cabab25)
J
Jirka Borovec committed
4921af29795ac6b53a437edffad8a07de7e67bbf
Parent: 4d84e51
Committed by Luca Antiga <[email protected]>
on 8/29/2025, 10:07:57 AM