huggingface/transformers

Pull Requests Commits

Stop using float64 in embedding math

Rocketknight1 committed 58 days ago

c5f100cf

Speed up modular conversion by reusing process pool (#46438)

guarin committed 58 days ago

Verified cc041637

revamp cv docs and fix rf-detr (#46219)

merveenoyan committed 58 days ago

Verified 1b7bc25c

Update hub metadata (#46379)

zucchini-nlp committed 58 days ago

Verified 1d185d43

extend DeepseekV4FlashIntegrationTest to non-cuda device (#46517)

sywangyi committed 58 days ago

Verified e3144397

[docs] deepgemm (#46361)

stevhliu committed 59 days ago

Verified 55115549

Fix `{image/video/audio}_token_ids` in `ProcessorMixin` (#46500)

hmellor committed 59 days ago

Verified d6a82ba8

[fix] regression introduced by #45534 (#46456)

eustlb committed 59 days ago

Verified b6bad752

[docs] tp for continuous batching (#46019)

stevhliu committed 59 days ago

Verified 153d6c16

Use torchvision's native LANCZOS interpolation instead of PIL fallback (#46496)

NicolasHug committed 59 days ago

Verified ff331709

skip quant cache tests for deepseek v4 (#46362)

kaixuanliu committed 59 days ago

Verified e569f888

revisit history parallel beam search tests to avoid unnecessary fix (#46495)

kaixuanliu committed 59 days ago

Verified 81fd7ae4

Add debugging info in `pr-ci-caller.yml` (#46505)

ydshieh committed 59 days ago

Verified e0779e46

Fix tests: 'Cohere2MoeModel' object has no attribute 'hf_device_map' (#46337)

kaixuanliu committed 59 days ago

Verified bfaf9216

Triton finegrained fp8/fp4 (#46407)

IlyasMoutawwakil committed 59 days ago

Verified 0c9e543f

Fix dtype mismatch in NemotronH/Zamba2 Mamba2 CUDA-kernel path (`out_proj`) (#46487)

yuekaizhang committed 59 days ago

Verified 7ac99129

Bump the actions group across 1 directory with 19 updates (#46414)

dependabot[bot] committed 59 days ago

Verified ceb27422

fix(falcon_mamba): recommend `pip install kernels` in fast-path warning (#46343)

Anai-Guo committed 59 days ago

Verified fb9f5b43

Log some information in `.github/workflows/pr-ci-post-dashboard-link.yml` (#46499)

ydshieh committed 59 days ago

Verified a635e94e

feat(quantizers): support non-weight param names in TorchAo safetensors loading (#46325)

agesf committed 59 days ago

Verified c54b69a1

docs: fix typo in make_list_of_images docstring (#46469)

ramkumar27072006 committed 59 days ago

Verified cbb65a48

fix: convert to fp32 first when doing image resize (#46262)

kaixuanliu committed 59 days ago

Verified 11f5a0ec

add XPU expectation for deepseek_ocr2 model tests (#46492)

kaixuanliu committed 59 days ago

Verified 5d82aacc

Fix sapiens2 tests: add XPU device expectations (#46488)

kaixuanliu committed 59 days ago

Verified 54b4108c

Add vLLM smoke test to CI (#46383)

hmellor committed 59 days ago

Verified e86792e3

Fix: interleaved RoPE application for MLA and Support Index Cache DSA indexer skip-topk sharing for GLM5 (#46372)

Dovis01 committed 59 days ago

Verified 51616377

extend deepseek v4 test to xpu (#46366)

sywangyi committed 62 days ago

Verified effde209

Quantization for small models (#46449)

sgerrard committed 62 days ago

Verified a921b4d8

Added cosmos3 model (#46146)

MaciejBalaNV committed 62 days ago

Verified f677e3a6

ci: fix slack report job — use AWS runner and authenticated GitHub API calls (#46448)

ydshieh committed 62 days ago

Verified b1ac5349

Older