huggingface/transformers

Pull Requests Commits

Fix PR CI workflow cancellation condition (#46276)

ydshieh committed 79 days ago

Verified 96d3ecce

[fix] toctree (#46106)

stevhliu committed 79 days ago

Verified 7b89db28

add more generic support for distributed trainer tests (#46109)

kaixuanliu committed 79 days ago

Verified c5292db6

add XPU Expectations for florence2 and lfm2_vl model test (#46275)

kaixuanliu committed 79 days ago

Verified da78b36f

Fix model parallel issue for altclip model and ChineseClip model (#45487)

kaixuanliu committed 79 days ago

Verified 08f13097

feat: Add support for JetBrains' `Mellum` v2 code generation model (#46112)

shadeMe committed 79 days ago

Verified 9cfa266f

Fix `StaticCache` building an empty layer list when `num_kv_shared_layers == 0` (#46235)

tengomucho committed 79 days ago

Verified 0ff380e0

Model parallel fix (#46230)

kaixuanliu committed 79 days ago

Verified e120f7ea

Fix TAPAS tokenizer crash on pandas 3.x string-dtype tables (#45916)

Rocketknight1 committed 79 days ago

Verified 1656d90b

Enable kernels-community/metal-flash-sdpa on MPS (#45974)

ArthurZucker committed 80 days ago

Verified bdca59aa

Fix inverted assertion in remove_handler (#46227)

SebTardif committed 80 days ago

Verified a8a01e53

[ShieldGemma2] Support attn_implementation dispatch (#46069)

YangKai0616 committed 80 days ago

Verified ec06284c

[Gemma4] Replace one-hot matmul with F.embedding in position embeddings (#46176)

Sriniketh24 committed 80 days ago

Verified bc8f70a9

fix: kosmos2.5: properly expand embeddings table (#45835)

nunq committed 80 days ago

Verified fd1d5b7d

find pytest launch error in torch 2.13.0.dev20260526 (#46252)

sywangyi committed 80 days ago

Verified f336a53d

Update a test to avoid writing to the default xet cache (#46250)

ydshieh committed 80 days ago

Verified 2d31b762

[Test][Kosmos2.5] Add XPU expectations for integration tests (#46135)

YangKai0616 committed 80 days ago

Verified 88070eff

[`Revert`] FSDP+Dtensor refactor related changes (#46246)

vasqu committed 80 days ago

Verified 295cee3e

Support FA2 flash_attn_with_kvcache for XPU continuous batching (#46028)

YangKai0616 committed 80 days ago

Verified a65bf6c9

[`Configs`] Fix layer type validation to include its mlp counterpart (#46220)

vasqu committed 80 days ago

Verified 70257e9a

Fix `num_items_in_batch` over-counting for causal LM losses (#46204)

qgallouedec committed 80 days ago

Verified 67265ef5

RF-DETR doc fixes (#46244)

merveenoyan committed 80 days ago

Verified 5adca7bf

Use `main` instead of commit SHA for now (#46241)

ydshieh committed 81 days ago

Verified 1a5a5a71

Enable push event (to main) for PR CI workflow (#46240)

ydshieh committed 81 days ago

Verified 5298d0ca

fix(hrm_text): Add XPU Expectations for tests (#46214)

kaixuanliu committed 81 days ago

Verified a6cd3b4e

Fix model parallel bugs for Gemma4 (#45817)

kaixuanliu committed 81 days ago

Verified f39b5c8b

Fix bnb 4bit/8bit quantization drop chunked tensors bug (#46210)

kaixuanliu committed 81 days ago

Verified 212863bc

[deepseek_v4] keep hc_head / sinks / position_bias in fp32 (#46198)

ArthurZucker committed 81 days ago

Verified 9ded3dbb

init FSDP through from_pretrained (#46102)

3outeille committed 81 days ago

Verified 0588858f

Fix FSDP2 and distributed checkpointing imports for older PyTorch versions (#46141)

ryota-komatsu committed 81 days ago

Verified 634500b9

Newer Older