huggingface/transformers

Pull Requests Commits

DTensor-based TP + FSDP2 shard-on-read composability

3outeille committed 2 days ago

607cc114

from_pretrained distributed refactor (FSDP2 + TP) (#44996)

3outeille committed 2 days ago

Verified 187ee5d4

3outeille committed 2 days ago

dfdfbdd9

Merge fsdp-vs-ddp (ruff fixes) into fsdp-core-model-loading

3outeille committed 3 days ago

b7c3dc5c

style: ruff lint + format fixes

3outeille committed 3 days ago

46e0fbe9

Merge fsdp-vs-ddp into fsdp-core-model-loading

3outeille committed 3 days ago

c9f30d9f

Merge distributed_api into fsdp-vs-ddp

3outeille committed 3 days ago

f4fe3304

Merge branch 'main' into distributed_api

3outeille committed 3 days ago

Verified 69bc48e1

fix: remove Copied from comments between @torch.jit.script and def for Python 3.13 compat (#44986)

Krishnachaitanyakc committed 3 days ago

Verified 2f624917

More small vllm fixes (#44990)

ArthurZucker committed 3 days ago

Verified 0e1978c9

fix(models): Fix Perceiver interpolate_pos_encoding interpolating to the source size (#44899)

harshaljanjani committed 3 days ago

Verified 2620c4dd

Allow `mm_token_type` be non-padded lists (#44563)

zucchini-nlp committed 3 days ago

Verified c532659b

3outeille committed 4 days ago

7c843391

Fix CPU 16 bytes alignment issue using equivalent fallback (#44970)

IlyasMoutawwakil committed 4 days ago

Verified c3b41c6c

refactor: unify QA calls (#44879)

tarekziade committed 4 days ago

Verified b3d7942f

Fix tie_word_embedding issues with `Qwen2VL` (#44976)

hmellor committed 4 days ago

Verified 28af8184

Support Modular (!!) + Configs in `check_auto_docstrings` (#44803)

yonigozlan committed 4 days ago

Verified 28e1cc59

[ `vllm x v5`] nit (#44971)

ArthurZucker committed 4 days ago

Verified 692d187d

LwDetrImageLoss: Fix dtype casting to prevent crash when using amp on cuda device (#44886)

m-matthias committed 4 days ago

Verified 18f88de7

[AMD CI] Gemma3/Gemma3n Expectations (#44972)

Abdennacer-Badaoui committed 4 days ago

Verified b94fee04

Refactor core_model_loading to support FSDP shard-on-read loading

3outeille committed 4 days ago

4b2a9216

Officially launch parse_response (#44674)

Rocketknight1 committed 4 days ago

Verified 15b3b892

fix load_best_model_checkpoint_at_end do not load the best model chec… (#44583)

wilnn committed 4 days ago

Verified e22f3088

fix: split MXFP4 dependency checks for specific error messages (#44930)

javierdejesusda committed 4 days ago

Verified 82106705

Fix failing `T5ModelIntegrationTest` (#44934)

Sai-Suraj-27 committed 4 days ago

Verified a8732d55

Config kwargs (#44953)

zucchini-nlp committed 4 days ago

Verified d4895f08

Fix: Update optimization.py (#44909)

anshuS1310 committed 4 days ago

Verified 2f121eaf

[CB] [Minor] Simplify test suite (#44858)

remi-or committed 4 days ago

Verified 4f8de834

Allow arbitrary template kwargs in processors (#44881)

zucchini-nlp committed 4 days ago

Verified 9ef15d70

Fix missing post_processor in DebertaV2Tokenizer causing no special t… (#44570)

umbilnm committed 5 days ago

Verified 6a056a16

Older