Pretrain, finetune ANY AI model of ANY size on 1 or 10,000+ GPUs with zero code changes.
bump: `bitsandbytes >=0.45.2,<0.47.0` & compatibility patch for `bnb.functional.int8_double_quant` (#20911)
* compatibility patch for `bnb.functional.int8_double_quant` * bitsandbytes >=0.45.2,<0.47.0 --------- Co-authored-by: Copilot <[email protected]> (cherry picked from commit c03660a8f84c7c1a09f3da230061b88e84430111)
J
Jirka Borovec committed
cf412834d06e5fc13ec0fe54a6246ea2d76e49ad
Parent: 7659176
Committed by Jirka B <[email protected]>
on 6/19/2025, 10:17:24 AM