huggingface/transformers

Pull Requests Commits

Fix Pyright type narrowing: assert EncoderDecoderCache in encoder-decoder branch

ydshieh committed 45 days ago

347b9fc9

ydshieh committed 46 days ago

892f3558

Scope fix to encoder-decoder only; restore elif prefill_chunk_size for decoder-only

ydshieh committed 47 days ago

f544b291

Clarify comment: replace 'folded in' with plain language

ydshieh committed 47 days ago

564e3fa6

Fix issue reference: #46421 is the issue, #46446 is the PR that fixed it

ydshieh committed 47 days ago

a611fca9

Clarify comment: chunked-prefill block folded in, not specific to that case

ydshieh committed 47 days ago

0e9de382

[generate] Fix CUDAGraphs tensor overwrite for encoder-decoder static cache

ydshieh committed 48 days ago

7d2a05a8

[TableTransformer, PI0] Fix stale expected values and OOM in integration tests (#48198)

ydshieh committed 48 days ago

Verified d56c55bf

Post two CI badges on a PR: CPU PR CI and GPU run-slow (#48190)

tarekziade committed 48 days ago

Verified e453228e

Fix stale expected values in integration tests (cuda sm_86 / Aug04 regressions) (#48171)

ydshieh committed 48 days ago

Verified 3ad85ca4

Assign a reviewer even when a codeowner has left, and route models by modality (#48085)

tarekziade committed 48 days ago

Verified e277d1ed

[VITS] Un-skip test_model_forward (#46375)

blipbyte committed 48 days ago

Verified bb593bfb

Apply context parallelism to the evaluation path (#48167)

qgallouedec committed 48 days ago

Verified 695bd682

Fix `gpt_oss` runs on GPU (#48118)

tarekziade committed 48 days ago

Verified 275b5245

🚨 TP dtensor API inference + training (#47579)

3outeille committed 48 days ago

Verified 861f4c41

[EsmFold2] Fix stale expected distogram logit values (#48182)

ydshieh committed 48 days ago

Verified d4c297d1

Fix Apr 05 integration test regressions (cuda sm_86) (#48170)

ydshieh committed 48 days ago

Verified cc33c370

Add MLU support to is_flash_linear_attention_available (#46995)

zsl223 committed 48 days ago

Verified 6a9e2d37

Fix ROCm SDPA-flash skip guard that crashes on RDNA GPUs (#47965)

Abdennacer-Badaoui committed 48 days ago

Verified 3cb4f80f

Fix integration test expected values for cuda sm_86 (Mar 15 regressions) (#48168)

ydshieh committed 48 days ago

Verified 251c2d4f

Fix DynamicCache reconstruction during ExecuTorch export (#47900)

eladsegal committed 48 days ago

Verified 3cd1a230

Revert "Support per-layer cache configuration and attention-mask selection" (#48175)

Cyrilvallez committed 48 days ago

Verified 6b3a9de1

Support per-layer cache configuration and attention-mask selection (#47901)

eladsegal committed 48 days ago

Verified c84ac341

[LLaVA] Fix pixtral integration tests for cuda sm_86 (#48166)

ydshieh committed 48 days ago

Verified 945dac91

Port ESMC and ESMFold2 to Transformers (#46419)

Rocketknight1 committed 49 days ago

Verified c8df3fc9

[Qwen2.5-Omni] Update stale expected values for cuda sm_86 (#48164)

ydshieh committed 49 days ago

Verified cdea840f

[Mistral3] Fix batched integration tests: padding_side=left + update expected values (#48161)

ydshieh committed 49 days ago

Verified dd6c64b7

Fix DeepSeek V2 default vocab size (#48159)

hmellor committed 49 days ago

Verified e4dd56a3

[InternVL] Fix stale expected values for Llama integration tests (cuda sm_80) (#48153)

ydshieh committed 49 days ago

Verified 0c928118

Retry transient network errors (RemoteDisconnected) in github_utils (#48124)

ydshieh committed 49 days ago

Verified 95119511

Older