PR #46269 DIstributed branch base

DIstributed branch base #46269

3outeille wants to merge 13 commits into main from distributed

Revert "[`Revert`] FSDP+Dtensor refactor related changes (#46246)"

18a712b9

not unifying but still better than before

863e12ba

3outeille changed the title ~~DIstributed branch~~ DIstributed branch base 75 days ago

mellum now inherits tp and sp plan from qwen3_moe

f914e301

handle dense + sparse mixing for mellum model in SP plan

2d01eb3f

Merge branch 'main' of https://github.com/huggingface/transformers in…

d672665e

check-repository-consistency

fb95e264

3outeille force pushed from 00922bbf to fb95e264 57 days ago

check-code-quality

61d3cbd1

fix

c3d75af3

simplify fsdp (#46679)

a5401678

Merge branch 'main' into distributed

5845f3a3

Tp param level (#46290)

30972993

Merge branch 'main' into distributed

08911d5f

Merge branch 'main' into distributed

245a7f6c

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Milestone

No milestone