huggingface/transformers

Pull Requests Commits

[serge] Fix 3 integration tests for model `deepseek_vl` failing with `other` (other (3)) (#48536)

sergereview[bot] committed 7 days ago

Verified c1b80cf3

[Improvement] Rework the docstring and comments of mHC (#48888)

remi-or committed 7 days ago

Verified 230e6f32

Don't fail model loading when the accelerator can't report free memory (#48136)

studioego committed 7 days ago

Verified 36ca5c6a

Fix Idefics2 padding when the first example has no image (#48753)

Arnavsharma2 committed 7 days ago

Verified 2e39c410

[`feat`] Allow untying `hidden_states[-1]` from `last_hidden_state` via the model config (#48087)

tomaarsen committed 7 days ago

Verified 12ee21d1

Fix GLM rope_parameters dict shared mutation across test instances (#48895)

ydshieh committed 7 days ago

Verified 9d4ad4b7

Fix sliding window cache when assistant model calls generate (#48280)

VaggelisGian committed 7 days ago

Verified 12b6da1f

Fix DFlash sampled candidates losing the batch dimension (#48281)

VaggelisGian committed 7 days ago

Verified af739912

[CB] Add pause mechanism (#48462)

remi-or committed 7 days ago

Verified 8e608337

Fix Pixtral image processor do_pad option (#48872)

jzakrzew committed 7 days ago

Verified 5a471cc1

Fix Minimax M2 partial_rotary_factor by remapping legacy rotary_dim (#48486)

dajiaohuang committed 7 days ago

Verified da9d2c46

Fix `StaticCache` for Mllama and enable `torch.compile` (#48141)

jiqing-feng committed 7 days ago

Verified ffddd251

Fix Glm4MoeIntegrationTest: split into 3 tests, switch to GLM-4.5-Air (#48820)

ydshieh committed 7 days ago

Verified dbea1093

fix(ci): harden GitHub Actions workflows (#48160) (#48834)

hf-security-analysis[bot] committed 7 days ago

Verified b22fe462

Fix flaky RfDetr test_save_load (#48869)

ydshieh committed 7 days ago

Verified af980eef

Fix ESMFold2 ligand iPLDDT weighting: mol_type non-polymer code is 3, not 4 (#48831)

faustomilletari committed 8 days ago

Verified 22e88feb

Add test for causal only variant of some encoder-decoder models (#48760)

nandan2003 committed 8 days ago

Verified c9d531dd

Fix kosmos (#48853)

Cyrilvallez committed 8 days ago

Verified 947d6b4a

[generate] Make all methods and logits processors agnostic to lm_head output size (#48846)

Cyrilvallez committed 8 days ago

Verified b180f143

Fix `reset` on the dynamic cache layers (#48809)

jiqing-feng committed 8 days ago

Verified a5d98c34

QA: deactivate rule 41 (#48852)

tarekziade committed 8 days ago

Verified 031cfc58

[GPTNeoXJapanese] Fix RoPE ignoring partial_rotary_factor (#48652)

blipbyte committed 8 days ago

Verified 2d92ded1

docs: fix broken #combining-with-fsdp2 anchor in expert_parallelism.md (#48854)

ydshieh committed 8 days ago

Verified 51b330bb

Qwen3.8 GGUF (#48660)

SunMarc committed 8 days ago

Verified 41f519e7

[CI] Add link checker (#48160)

stevhliu committed 8 days ago

Verified e9af3a32

🚨 🚨Bring some dinos to modern standards (#46266)

molbap committed 8 days ago

Verified f2f6074b

better distinguish masked/all-to-all path for nicer API

3outeille committed 8 days ago

b2b0d324

raise in `resolve_parallel_plans` when an override key matches no module / parameter / existing plan key

3outeille committed 8 days ago

2dc99524

[apply_chat_template] pass sampling_rate to call (#48794)

eustlb committed 8 days ago

Verified 3dec062f

[docs] gguf (#46357)

stevhliu committed 8 days ago

Verified 626e7ff9

Newer Older