Flash mla interface #44054
ArthurZucker
wants to merge
23 commits into
main
from
flash-mla-interface
init
73cf287e
format
b4c0a929
not indexer_rope_interleave
85c3bd1a
set MLA rope interleave to False
545d91ac
get rid of interleave in apply_rotary_pos_emb
6d0f6a87
Merge branch 'main' into glm-dsa
d829f860
reintroduce attention interface
e84b43c9
reset _cached_keys
558989a0
remove yarn
62454529
fix tp plan for multi node runs
299c53c3
tentatively add flash-mla
125f9949
Merge branch 'main' of github.com:huggingface/transformers into flash…
776378b9
skip more tests
4f89abe5
LalithaMV
approved these changes
on 2026-02-16
LalithaMV
requested changes
on 2026-02-16
Merge branch 'main' into flash-mla-interface
71e10bd4
fuck the tp plan is wrong
9e7a2e19
small fixes
1dff69e3
Merge branch 'main' into flash-mla-interface
80cf8850
yup
7b3ba9a5
Merge branch 'main' of github.com:huggingface/transformers into flash…
c32bdea8
Merge branch 'main' of github.com:huggingface/transformers into flash…
d8f79e43
current changes
309c8fd7
nit
1f40e8bb
Merge branch 'flash-mla-interface' of github.com:huggingface/transfor…
e0e407f1
Login to write a write a comment.
Login via GitHub