huggingface/transformers

Pull Requests Commits

Merge branch 'main' into fix-mamba2-decode-squeeze

Cyrilvallez committed 77 days ago

Verified b5605709

Fix TP inference for tied embedding (#47503)

3outeille committed 78 days ago

Verified 279e6efc

Fix Mamba2-family decode step: restore the seq-dim squeeze

qgallouedec committed 80 days ago

4b1f46f1

Add AXK2 from SKT (#47528)

vasqu committed 80 days ago

Verified b6d5084f

Fix `causal_conv1d_fn` positional `activation` colliding with hub kernel's `seq_idx` (#47527)

qgallouedec committed 80 days ago

Verified 77a78c9f

Fix failing tests for mimo_v2_flash (#47284)

kaixuanliu committed 80 days ago

Verified ffec245b

Refactor all linear attention models to latest best standards for convolution (#47452)

Cyrilvallez committed 80 days ago

Verified a79421ac

Make tokenization_mistral_common importable without mistral_common installed (#47397)

juliendenize committed 80 days ago

Verified 98a1e7d8

Use --flake-runs=1 in check_bad_commit.py for PR comment CI (#47522)

ydshieh committed 80 days ago

Verified 5b32a41f

Deprecate the old response_schema (#47320)

Rocketknight1 committed 80 days ago

Verified 77bcb281

fix: add pickle support to _LazyConfigMapping for spawn multiprocessing (#46026)

kfojcik-intel committed 80 days ago

Verified f07253a9

Fix potential ReDoS by escaping tokenizer filename used as regex pattern (#47498)

hameedibrh committed 80 days ago

Verified 7e0ff853

Fix vision position-embedding init width fallback in Phi4Multimodal (#47509)

yuvrajnode committed 80 days ago

Verified 0f2f0518

Delete old deprecations (#47518)

zucchini-nlp committed 80 days ago

Verified 54c560ac

Run only @slow tests in PR comment CI (#47521)

ydshieh committed 80 days ago

Verified 42da4a71

Fix incorrect type hint (#47519)

hmellor committed 80 days ago

Verified d014dd2c

Deprecate CB config in gen configuration (#47291)

remi-or committed 80 days ago

Verified b98026a3

Allow metal-flash-sdpa for OpenAIPrivacyFilter on MPS (#46740)

ArthurZucker committed 81 days ago

Verified 4b57c08d

[Offloading] [Bugfix] Fix fully offloaded model saving (#47336)

kylesayrs committed 81 days ago

Verified e9c987a3

CI: fix torchaudio pinning +proper break in rnnt (#47422)

tarekziade committed 81 days ago

Verified 73f4d225

add_axk1 (#46867)

kmswin1 committed 81 days ago

Verified c27860a2

Fix Qwen2.5-Omni Token2Wav DiT rotary embedding layout (interleaved cos/sin) (#47403)

HenryVarro666 committed 81 days ago

Verified 363b2a9f

CI: add reproduce mode to serge verify caller (#47492)

tarekziade committed 81 days ago

Verified 00d795b4

feat[vLLM x v5]: Make audio optional and support multiple audios in VibeVoice ASR processor (#47483)

harshaljanjani committed 81 days ago

Verified 720ae6e6

[fix][whisper]: fix max_new_tokens handling (#46795)

eustlb committed 81 days ago

Verified ee48a631

Isolate MLA KV expansion to make it easier to bypass (#47460)

hmellor committed 81 days ago

Verified 6b0f8227

Update serve chat parsing (#46267)

SunMarc committed 81 days ago

Verified 12f2d155

Fix loss alignment and Trainer token counting for encoder decoder models (#46903)

OmkumarSolanki committed 81 days ago

Verified 0c1829dd

Fix Gemma4 audio feature dtype mismatch in masked_scatter (#47482)

danielhanchen committed 81 days ago

Verified bc14d7e6

Fix the HunyuanVL's torchvision backend (#47499)

Mi-Jiazhi committed 81 days ago

Verified 60570c1f

Older