huggingface/transformers

Pull Requests Commits

[docs] FlashAttention kernel fallback (#47345)

stevhliu committed 5 hours ago

Verified 28596623

Fix inkling feature extractor (#47349)

ArthurZucker committed 5 hours ago

Verified d7d894cf

Small improvement (#47348)

Cyrilvallez committed 7 hours ago

Verified 0a408c33

Add Inkling model (#47347)

molbap committed 7 hours ago

Verified c375c5e8

🚨 Pass tp_plan from lm_heads (#47253)

molbap committed 7 hours ago

Verified f5a67a46

ci: cover xet as well (runtime error) (#47338)

tarekziade committed 9 hours ago

Verified cd7456ea

[docs] TokenizersBackend fallback (#47302)

stevhliu committed 10 hours ago

Verified 90457957

Pin kernels to compatible version in benchmark workflow (#47339)

tarekziade committed 13 hours ago

Verified e52d0fd6

Fix models for which we don't have a dedicated tokenizer class, and the listed one is incorrect (#45936)

itazap committed 17 hours ago

Verified e194978a

Resolve continuous batching XPU availability checks at runtime (#47185)

kaixuanliu committed 20 hours ago

Verified 498d6e98

[Nit] Add kernels_fallback_ok kwarg to is_flash_attn_N_available (#47318)

remi-or committed 21 hours ago

Verified 4dd59f71

[Nit] Add expectations for gemma4 tests on H100 (#47311)

remi-or committed 23 hours ago

Verified b7f01015

[docs] DeepGEMM requirements (#47324)

stevhliu committed 1 day ago

Verified ab1771c9

:rotating_light: Fix GPTBigCode and GPTNeoX for the Transformers modelling backend for vLLM (#47198)

hmellor committed 1 day ago

Verified 70b1ee26

DeepGEMM shouldn't pad on SM90 (#47313)

IlyasMoutawwakil committed 1 day ago

Verified a9e26099

Fix half-precision torch.compile crash in DETR-family sine position embeddings (#47238)

David-Wu1119 committed 1 day ago

Verified 7c8b1d44

Fix hardcoded paths in siglip checkpoint/vocab loading (#47178)

XanxusCrypto committed 1 day ago

Verified a6895655

[generation] Fix misleading synced_gpus warning in continuous batching (#47158)

Partha-Shankar committed 1 day ago

Verified f1a2a70f

Update AMD CI runner groups to amd-mi300 (#47307)

Abdennacer-Badaoui committed 1 day ago

Verified fc7cf967

Point to Gemma 4 model in Gemma4ForCausalLM docstring example (#47255)

lefft committed 2 days ago

Verified 6805d433

Fix Qwen Omni batched text postprocessing (#47197)

Sunt-ing committed 2 days ago

Verified 4a272654

Fix AqlmConfig error messages to say "int" instead of "float" (#47089)

Sreekant13 committed 2 days ago

Verified 81eb7c3b

Fix check for interactive stdout in _style function (#47283)

smart8986 committed 2 days ago

Verified 11ed2ff4

Fix get_json_schema crash on non-string docstring choices (#47072)

Sreekant13 committed 2 days ago

Verified 529858a5

Make `MODEL_IDS_TO_TOKENIZERS_BACKEND` capture all DeepSeek R1 distills (#47296)

hmellor committed 2 days ago

Verified 35bb43e7

Update doc preprocessing regex to prevent ReDoS (#47187)

WilliamRoyNelson committed 2 days ago

Verified 13dfd790

[generate] Add proper MTP support (#46229)

Cyrilvallez committed 2 days ago

Verified 01dd2fd8

Shard on read Dtensor aware (#46717)

3outeille committed 2 days ago

Verified 4517f141

Switch AMD daily CI to mi300 runners (#47259)

Abdennacer-Badaoui committed 4 days ago

Verified 63f32a87

tests: reduce processor test memory usage by using tiny Hub checkpoints (#47213)

ydshieh committed 5 days ago

Verified 1f8daee0

Older