huggingface/transformers

Pull Requests Commits

Merge remote-tracking branch 'origin/fix-continuous-batching-device-mesh-import' into fix-continuous-batching-device-mesh-import

SunMarc committed 5 days ago

638f222f

Avoid eager torch.distributed.tensor.device_mesh import in continuous_batching

SunMarc committed 5 days ago

c9161099

Guard torch.distributed.tensor.device_mesh import in continuous_batching

SunMarc committed 5 days ago

0201b877

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

ydshieh committed 7 days ago

Verified 1a5a5a71

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

ydshieh committed 7 days ago

Verified 5298d0ca

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

kaixuanliu committed 7 days ago

Verified a6cd3b4e

Fix model parallel bugs for Gemma4 (#45817)

kaixuanliu committed 7 days ago

Verified f39b5c8b

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

kaixuanliu committed 7 days ago

Verified 212863bc

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

ArthurZucker committed 8 days ago

Verified 9ded3dbb

init FSDP through from_pretrained (#46102)

3outeille committed 8 days ago

Verified 0588858f

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

ryota-komatsu committed 8 days ago

Verified 634500b9

Fix Gemma4 Array Mask Indexing (#46203)

petecao committed 8 days ago

Verified 5a4b70f8

utils: handle flash_attn missing from importlib packages_distributions without crashing (#45524)

SAY-5 committed 8 days ago

Verified 02af5bfa

fix model parallel device mismatch issue in `create_bidirectional_mask` (#46221)

kaixuanliu committed 8 days ago

Verified 710742f1

Fix a regression in encoder-decoder generation cache initialization (#46111)

kaixuanliu committed 8 days ago

Verified 90e3c4fa

[AMD CI] revert AMD mi325 hf-workflows ref from SHA back to @main (#46213)

Abdennacer-Badaoui committed 8 days ago

Verified 6db6fdb9

Trainer.compute_loss: fix loss over-counting under TP and EP-as-TP (#45994)

AmineDiro committed 8 days ago

Verified b131ec14

[GLM-4.6V] Update with GLM-GA Processor (#46184)

zRzRzRzRzRzRzR committed 8 days ago

Verified 89c0c2ed

🚨 [ALM] Add base model without head (#45534)

eustlb committed 9 days ago

Verified 027d1a97

update xpu expectation for falcon mamba (#46086)

sywangyi committed 9 days ago

Verified 08692e3c

chore: enable Dependabot weekly GitHub Actions bumps (#46157)

hf-dependantbot-rollout[bot] committed 9 days ago

Verified 9b1bf032

Fix Gemma4 use_bidirectional_attention="all" mask behavior (#46079)

oliverholworthy committed 9 days ago

Verified ceb7ba69

Fix loading with only 1 device or distributed config (#46197)

Cyrilvallez committed 9 days ago

Verified 4a2e3652

Fix TypeError on list-typed ignore_keys_at_rope_validation in RoPE config (#46142)

Charly21r committed 9 days ago

Verified ece1ea06

Support XPU autocast dtype fallback for FlashAttention (#46199)

YangKai0616 committed 9 days ago

Verified 7bc093b7

Fix path traversal when saving named chat templates (#46191)

LinZiyuu committed 9 days ago

Verified eaaaf849

Fix is_last off-by-one in MaskGenerationPipeline for partial batches (#46136)

J3r3myPerera committed 9 days ago

Verified 47949d3a

Fix wrong variable in check_model_type isinstance check (#46080)

SebTardif committed 9 days ago

Verified 118ddfb6

Enable passing kwargs through RoFormer models (#46171)

ir2718 committed 9 days ago

Verified 0b2da85d

Update cohere2_moe tp_plan (#46189)

Cyrilvallez committed 10 days ago

Verified e65c3a2d

Older