[Model] Lfm2Moe #41401

Cyrilvallez merged 30 commits into huggingface:main from paulpak58:lfm2_moe
paulpak58
paulpak58 [new-models] LFM2-MoE
d3e71a9b
paulpak58 [docs] add in template lfm2_moe doc files
ee096db6
paulpak58 [configuration] update configuration class
1ef7c59b
paulpak58 [modular][lfm] minor: fix rotary_emb typo
d42252a4
paulpak58 [modeling] modular/modeling files for Lfm2Moe
79b3f20d
paulpak58 [modeling][lfm2_moe] fix Lfm2Moe modular/modeling
5bc8e0d1
paulpak58 [configuration][lfm2_moe] update configuration keys with latest confi…
bfd559a1
paulpak58 [misc] make fixup
e410997f
paulpak58 [modular][lfm2_moe] address comments: dtype, mlp, buffers
7888413a
paulpak58 [configuration][lfm2_moe] add initializer_range
d048b127
paulpak58 [modular][lfm2_moe] include init_weights to pass test_initialization
c0486222
paulpak58 [tests][causal_lm] include pos_emb as possible rope attribute
d10a8c99
paulpak58 [modeling][lfm2_moe] remove load_balancing_loss_func due to lack of s…
7114e62c
paulpak58 [misc] make style
1b74ceea
paulpak58 [modeling][lfm2_moe] MoE refactor PR update in LFM2Moe
fb65ff5b
paulpak58 [tests] lfm2_moe: unit tests
f4785d93
paulpak58 [misc] update LFM2-8B-A1B repo id
dd63210c
paulpak58 [tests] lfm2: update ModelTests for lfm2
fc7d2eaf
mlabonne Update LFM2 documentation
daa35aa0
mlabonne Add Lfm2Moe documentation
96fbd87e
paulpak58 [misc] fix ci
c7bac38e
paulpak58 [docs] remove trust_remote_code
37d2ffff
paulpak58 [misc] ci: fix modular
23c2c991
Cyrilvallez Merge branch 'main' into lfm2_moe
1427f48b
Cyrilvallez reapply modular
8b6aa838
Cyrilvallez simplify
cd873d4b
Cyrilvallez
github-actions
Cyrilvallez remove static address and inplace op
0c67e26d
Cyrilvallez simplify
44ff5d25
github-actions
Cyrilvallez simplify a bit more the modular
38130f5e
Cyrilvallez imports
b3a0924e
Cyrilvallez
Cyrilvallez approved these changes on 2025-10-07
Cyrilvallez Cyrilvallez merged 0c9a72e4 into main 1 year ago

Login to write a write a comment.

Login via GitHub

Reviewers
Assignees
No one assigned
Labels
Milestone