Pretrain, finetune ANY AI model of ANY size on 1 or 10,000+ GPUs with zero code changes.
Fix TQDM progress bar showing the wrong total when using a finite and iterable dataloader (#21147)
* fix implementation reset * add testing * changelog
N
Nicki Skafte Detlefsen committed
14a57c76bd4fe9b7fd8c1a313441a514b41563e4
Parent: 630db82
Committed by GitHub <[email protected]>
on 9/2/2025, 8:05:22 AM