huggingface/transformers

Pull Requests Commits

Restore TokenizersBackend dispatch for MODELS_WITH_INCORRECT_HUB_TOKENIZER_CLASS

ArthurZucker committed 53 days ago

12c8a3cd

Fix pageable H2D copies in Gated DeltaNet PyTorch fallback (#45665)

ruixiang63 committed 53 days ago

Verified ca72aa0a

Fix UnboundLocalError in shard_and_distribute_module for replicated parameters (#45675)

Abdennacer-Badaoui committed 53 days ago

Verified 4287660b

No serving in quality docker image (#45677)

ydshieh committed 53 days ago

Verified 1579083b

Laguna XS.2 implementation (#45673)

joerowell committed 53 days ago

Verified 2df23260

[MistralCommonBackend] Soften validation mode and apply_chat_template arguments check (#45628)

juliendenize committed 53 days ago

Verified 31469146

Fix `NameError: PeftConfigLike` triggered by `PreTrainedModel.__init_subclass__` (#45658)

qgallouedec committed 53 days ago

Verified 61461a7b

Fix cross-attention cache layer type for T5Gemma2 long inputs (#45540)

Beichen-Ma committed 53 days ago

Verified 63378d94

chore(typing): added modeling_utils to ty (#45425)

tarekziade committed 54 days ago

Verified 133d68e2

model: Add DEIMv2 to Transformers (#44339)

harshaljanjani committed 54 days ago

Verified 739b46cc

[Qwen3.5] Fix GDN linear attention multi-token cached forward (#45513)

kashif committed 54 days ago

Verified f53ca05d

[gemma4] infer from config instead of hardcoding (#45606)

eustlb committed 54 days ago

Verified 5d24d8c8

Update quants tests (#45480)

SunMarc committed 54 days ago

Verified e651c68e

Fix GraniteMoeHybrid _update_mamba_mask crash on attention-only models (#45514)

tianhaocui committed 54 days ago

Verified 617e752d

🔴🔴🔴 fix: skip `clean_up_tokenization` for BPE tokenizers in `PreTrainedTokenizerFast` (#44915)

maxsloef-goodfire committed 54 days ago

Verified bbb51c83

Fix colmodernvbert tests (#45652)

Cyrilvallez committed 54 days ago

Verified 7fdf09d3

[CB] [Major] Add CPU request offloading (#45184)

remi-or committed 54 days ago

Verified 5d755283

Fix peft constructors (#45622)

Cyrilvallez committed 54 days ago

Verified 545a591f

chore: speedup modular converter (~30%) (#45046)

tarekziade committed 54 days ago

Verified 99cdd438

Fix whisper return language (#42227)

FredHaa committed 54 days ago

Verified 4f85f854

Add `supports_gradient_checkpointing` to `NemotronHPreTrainedModel` (#45625)

sergiopaniego committed 54 days ago

Verified ded2b747

Raise clear error for `problem_type="single_label_classification"` with `num_labels=1` (#45611)

gaurav0107 committed 56 days ago

Verified c472755e

CircleCI with torch 2.11 (#45633)

ydshieh committed 57 days ago

Verified 678e871f

chore: bump doc-builder SHA for main doc build workflow (#45631)

rtrompier committed 57 days ago

Verified 622b8e95

Allow more artifacts to be download in CI (#45629)

ydshieh committed 57 days ago

Verified 23ca4372

chore(qa): split pipeline and add type checking (#45432)

tarekziade committed 57 days ago

Verified f0f456b3

Skip failing offloading tests (#45624)

Cyrilvallez committed 57 days ago

Verified a66638d8

generate: drop stale num_return_sequences warning on continuous batching path (#45582)

joaquinhuigomez committed 57 days ago

Verified f0a5a1cc

Remove unnecessary generate warnings (#45619)

Cyrilvallez committed 57 days ago

Verified 16f3dded

fix: compute auxiliary losses when denoising is disabled in D-FINE (#45601)

Abineshabee committed 57 days ago

Verified 5cf79514

Older