huggingface/transformers

Pull Requests Commits

Update deepseek v4 snippet: model id, quant config

pcuenca committed 71 days ago

6285d54d

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

ydshieh committed 71 days ago

Verified 2d31b762

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

YangKai0616 committed 71 days ago

Verified 88070eff

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

vasqu committed 71 days ago

Verified 295cee3e

Support FA2 flash_attn_with_kvcache for XPU continuous batching (#46028)

YangKai0616 committed 72 days ago

Verified a65bf6c9

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

vasqu committed 72 days ago

Verified 70257e9a

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

qgallouedec committed 72 days ago

Verified 67265ef5

RF-DETR doc fixes (#46244)

merveenoyan committed 72 days ago

Verified 5adca7bf

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

ydshieh committed 72 days ago

Verified 1a5a5a71

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

ydshieh committed 72 days ago

Verified 5298d0ca

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

kaixuanliu committed 72 days ago

Verified a6cd3b4e

Fix model parallel bugs for Gemma4 (#45817)

kaixuanliu committed 72 days ago

Verified f39b5c8b

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

kaixuanliu committed 72 days ago

Verified 212863bc

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

ArthurZucker committed 72 days ago

Verified 9ded3dbb

init FSDP through from_pretrained (#46102)

3outeille committed 73 days ago

Verified 0588858f

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

ryota-komatsu committed 73 days ago

Verified 634500b9

Fix Gemma4 Array Mask Indexing (#46203)

petecao committed 73 days ago

Verified 5a4b70f8

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

SAY-5 committed 73 days ago

Verified 02af5bfa

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

kaixuanliu committed 73 days ago

Verified 710742f1

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

kaixuanliu committed 73 days ago

Verified 90e3c4fa

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

Abdennacer-Badaoui committed 73 days ago

Verified 6db6fdb9

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

AmineDiro committed 73 days ago

Verified b131ec14

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

zRzRzRzRzRzRzR committed 73 days ago

Verified 89c0c2ed

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

eustlb committed 73 days ago

Verified 027d1a97

update xpu expectation for falcon mamba (#46086)

sywangyi committed 73 days ago

Verified 08692e3c

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

hf-dependantbot-rollout[bot] committed 73 days ago

Verified 9b1bf032

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

oliverholworthy committed 73 days ago

Verified ceb7ba69

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

Cyrilvallez committed 73 days ago

Verified 4a2e3652

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

Charly21r committed 74 days ago

Verified ece1ea06

Support XPU autocast dtype fallback for FlashAttention (#46199)

YangKai0616 committed 74 days ago

Verified 7bc093b7

Older