Pretrain, finetune ANY AI model of ANY size on 1 or 10,000+ GPUs with zero code changes.
Fix attribute error on `_NotYetLoadedTensor` after loading checkpoint into quantized model with `_lazy_load()` (#20121)
A
awaelchli committed
6c70dd7cf0db574b4b43f7b1fc07c3c8e7a83731
Parent: f91349c
Committed by GitHub <[email protected]>
on 7/24/2024, 9:39:40 AM