Pretrain, finetune ANY AI model of ANY size on 1 or 10,000+ GPUs with zero code changes.
Fix FSDP mixed precision semantics and add user warning (#21361)
D
Deependu committed
05ae8b37c9aed02d20b334f0e9f44fee6ed31b8d
Parent: 9a10959
Committed by GitHub <[email protected]>
on 11/21/2025, 12:17:04 PM