huggingface/transformers

Pull Requests Commits

oupsi limit case

Cyrilvallez committed 28 days ago

Verified 336119f5

Cyrilvallez committed 28 days ago

Verified 689ded41

Cyrilvallez committed 28 days ago

Verified a27245e2

[generate] Always correctly restrict assisted decoding with max length/eos token (#48981)

Cyrilvallez committed 28 days ago

Verified 87d34bfa

🚨 Remap indexers layer_type (#48974)

Cyrilvallez committed 28 days ago

Verified e2d83fdf

[generate] Simplify candidate generators by removing required `update_candidate_strategy` (#48982)

Cyrilvallez committed 28 days ago

Verified a2fc7527

Let `prefix_allowed_tokens_fn` override model `-inf` and raise an exception on unsatisfiable generation constraints. (#48927)

ksh108405 committed 29 days ago

Verified 66880ecc

Fix Qwen3OmniMoeIntegrationTest OOM (#48987)

ydshieh committed 29 days ago

Verified 5bb12bb4

Fix SwitchTransformers Top1 router: raw logits, expert capacity accounting, and router losses (#48421)

yurekami committed 29 days ago

Verified 13d21d5b

[CB] [Major] Upgrade the cache to support different attention types (#47809)

remi-or committed 29 days ago

Verified d94b0f7f

Fix vibevoice TTS batched audio index (#48902)

ebezzam committed 29 days ago

Verified a10365f9

Reduce peak memory in examples_torch CI job (OOM fix) (#48983)

ydshieh committed 29 days ago

Verified 21912efc

Fix typos in DeepseekV4 comments (#48953)

Janiarafath committed 29 days ago

Verified ae05b6b9

Contract the mamba2 chunk scan with einsum instead of broadcast-then-sum (#48978)

tarekziade committed 29 days ago

Verified a52f6593

Fix image processor class-level size mutation and min_pixels handling (#48916)

Gracy769 committed 29 days ago

Verified d53e5876

Add pose estimation keypoint preprocessing to Sapiens2ImageProcessor (#47199)

Sainava committed 29 days ago

Verified e86881ed

Better attn default for gguf (#48935)

SunMarc committed 29 days ago

Verified abd76d12

Skip an unnecessary image copy in the torchvision image normalization path (#48897)

jzakrzew committed 29 days ago

Verified 207fca7b

Fix more stuff (#48979)

zucchini-nlp committed 29 days ago

Verified c210bb39

Fix infeasible cost matrix errors in hugarian matcher losses (#47730)

guarin committed 29 days ago

Verified e138c4d5

Keep image processor backends in sync on keys and dtypes (#48739)

yupengtang committed 29 days ago

Verified a7d9d949

Resolve the Hub revision once per load instead of passing a private _commit_hash around (#47611)

Wauplin committed 29 days ago

Verified d67c7293

Fix AXK2 integration test: update CUDA expected text and rename class (#48941)

ydshieh committed 32 days ago

Verified c587bc88

docs: fix dead doc links in i18n READMEs and the xlnet docstring (#48905)

simpleqt committed 32 days ago

Verified 3713bd83

Align special tokens on the text config (#48847)

qgallouedec committed 32 days ago

Verified ea31b0c6

DOC: Improve documentation for remap_legacy_layer_types function (#48651)

mathewOracle committed 32 days ago

Verified 266cc923

docs: fix docstring parameters that do not match signatures (#48904)

simpleqt committed 32 days ago

Verified c9ad4a71

Fix tests due to dropping attn mask (#48903)

SunMarc committed 32 days ago

Verified b856a67b

Improve lazy import error messages (#48602)

karatarassul4-max committed 32 days ago

Verified 7301fbc6

fix(vibevoice-asr): use integer ceiling division for audio token count (#48864)

ege-arhan committed 32 days ago

Verified 719d8090

Older