[Feat] Add FP8 training support#758

Draft

fishcrap wants to merge 43 commits intomainfrom

Commits on Dec 2, 2025

add megatron training args
fishcrap
committed

Commits on Dec 3, 2025

Commits on Dec 4, 2025

Commits on Dec 5, 2025

Merge remote-tracking branch 'github/main' into sxj/fp8_train
fishcrap
committed

Commits on Dec 8, 2025

Commits on Dec 10, 2025

Commits on Dec 11, 2025

fix fp8 scale_inv and weight not in same bin
fishcrap
committed

Commits on Dec 17, 2025

Commits on Dec 22, 2025

Commits on Dec 23, 2025

Commits on Dec 24, 2025