huggingface/transformers

Pull Requests Commits

moe experts, router and sparse block canonicals + renames

itazap committed 1 day ago

51b0534f

breaking: mlp renames hitting peft defaults, legacy (bloom, gptj)

itazap committed 3 days ago

bc65f74d

breaking: mlp renames hitting peft defaults (gpt_neox, llama4)

itazap committed 3 days ago

ce08d660

mlp renames, legacy models

itazap committed 3 days ago

ac13a470

mlp canonicals and renames, non-legacy models

itazap committed 3 days ago

96a3c843

[`kernels`] Bump min version (#50509)

vasqu committed 3 days ago

Verified cfbca2d2

Fix flaky `test_eager_matches_batched_and_grouped_inference` in fp16 / bf16 (#50500)

ydshieh committed 3 days ago

Verified 4816e89f

Skip the query/key head repeat for grouped value heads in gated DeltaNet (#49141)

manav-simile committed 3 days ago

Verified 6096273e

Fix flaky backend equivalence tests with Pillow 13 (follow-up of #50105) (#50510)

ydshieh committed 3 days ago

Verified 70152d90

Add Jev + decision core (#50054)

ArthurZucker committed 3 days ago

Verified 31d73929

[FSDP tests] Fork CPU ranks from a forkserver instead of spawning them (#50444)

ydshieh committed 3 days ago

Verified 7a0a6fd6

fix(pipelines): respect audio classification problem type (#50329)

BlazeSR10hawk committed 3 days ago

Verified cc405bb3

fix: forward tool schemas to response parsers (#50385)

BlazeSR10hawk committed 3 days ago

Verified de058827

Doge: note that the MoE router split is not causal in a full forward pass (#50420)

Andreas-Hum committed 3 days ago

Verified 9c5bdb1d

Update the longrope inv_freq buffer in place (#50480)

hmellor committed 3 days ago

Verified b2556881

Retry the attention kernel load after a failed `lazy_import_flash_attention` (#50391)

albertvillanova committed 3 days ago

Verified 75ca9731

Keep the GLM-5.3-Flash mHC base and scale in float32 (#50413)

albertvillanova committed 3 days ago

Verified b6401d66

fix failed cases for beam model (#50235)

kaixuanliu committed 3 days ago

Verified bbbf45fb

[torch 2.15] Update qwen3_vl_moe FA2 integration test expected values (#50481)

ydshieh committed 3 days ago

Verified f262a8cd

fix(gemma3n): Clamp the AltUp magnitudes instead of using a CPU epsilon tensor (#50457)

harshaljanjani committed 3 days ago

Verified 431e9e9d

[Mistral Large 4] Fix flaky test_left_padding_compatibility (#50479)

ydshieh committed 4 days ago

Verified de35c2b1

[torch 2.15] Update integration test expected values for A10 (cuda 8.6) (#50478)

ydshieh committed 4 days ago

Verified 0f7a8768

[Docker] Update Nvidia CI and Flash Attn CI images to torch 2.15 / cu132 (#50467)

ydshieh committed 5 days ago

Verified a6f5c867

Clarify warmup_steps error message to warmup_steps >= 0 (#50364)

adi274903 committed 6 days ago

Verified 0f53828e

[Gemma4] Stop prefill at last KV-owning layer when logits_to_keep=1 (#49640)

pculliton committed 6 days ago

Verified 2c6a00f9

Fix duplicated docstring (#50446)

Cyrilvallez committed 6 days ago

Verified 073a202c

chore(ci): refresh workflow references (#50436) (#50437)

hf-security-analysis[bot] committed 6 days ago

Verified 04b0d424

[serge] Fix 2 integration tests for model `pegasus` failing with `other` (other (2)) (#49507)

sergereview[bot] committed 6 days ago

Verified d9246b06

[distributed - TPU] Fix tied-weight comparisons on TPUs when using TP (#50406)

tengomucho committed 6 days ago

Verified 92b21f8d

vLLM CI: fix install and KV-cache memory, add failure report job (#50436)

ydshieh committed 6 days ago

Verified 826edb70

Older