[litgpt benchmark] enable force_recompute_fp8_weight_in_bwd
when torchao.float8
is used with FSDP2
#1528
Azure Pipelines / lightning-thunder (ipynb) (jupyter ubuntu22.04 | cuda 12.1 | torch-nightly)
succeeded
Dec 17, 2024 in 9m 54s
jupyter ubuntu22.04 | cuda 12.1 | torch-nightly succeeded
Loading