huggingface/transformers

Pull Requests Commits

[core] Fix quark (#42457)

MekkCyber committed 160 days ago

Verified bebfab06

fix(Qwen3VLCausalLMOutputWithPast): missing `hidden_states` and `atte… (#42609)

casinca committed 160 days ago

Verified 0c3d043e

Every model forward() should have **kwargs (#42603)

Rocketknight1 committed 160 days ago

Verified 9b74e4c4

Update revision so that there is a safetensors model (#42618)

LysandreJik committed 160 days ago

Verified 3cdccba0

extend FA2 and other cases to XPU, (#42536)

yao-matrix committed 160 days ago

Verified 2e93004a

Fix some models cache initialization (#42586)

albertvillanova committed 161 days ago

Verified a48d68c6

Fix Ernie Moe Test (#42595)

vasqu committed 161 days ago

Verified 9e82c779

repo. consistency bot (#42575)

ydshieh committed 161 days ago

Verified d5d87934

Align tie weights in Idefics (#42551)

Cyrilvallez committed 161 days ago

Verified 81aabe72

Add SDPA support for PatchTST model (#42465)

Furkan-rgb committed 161 days ago

Verified dd6cfdd3

Allow `validation_fn` to be `None` in `validate_rope` (#42601)

hmellor committed 161 days ago

Verified b8d5018e

Add torch compile to CB (#42516)

remi-or committed 161 days ago

Verified ef780bf1

[Quantization] Remove dequant fp8 config (#42596)

MekkCyber committed 161 days ago

Verified 7266f50b

Fix the FA2 logic in the longcat_flash model (#42549)

YangKai0616 committed 161 days ago

Verified c0328af6

Fix mixed torch.Tensor and DTensor in generate when use FSDP2 + LoRA (#42436)

Xiao-Chenguang committed 161 days ago

Verified 17c7c496

fix : cast into floats AFTER all assignments (#42587)

marconaguib committed 161 days ago

Verified 130dc471

fix: correct typos in code comments (#42577)

arrdel committed 161 days ago

Verified 8ba286ba

fix the FSDP1 default for reshard_after_forward (#42578)

winglian committed 161 days ago

Verified cba7ae86

Fix loaded data order bug when resuming from epoch >= 1 (#40691)

ngazagna-qc committed 161 days ago

Verified 629c0da4

[XPU] Fix fp8 UT patch func (#42584)

YangKai0616 committed 161 days ago

Verified 75c135d7

Use `getattr` in `standardize_rope_params` because `rope_parameters` not always present (#42593)

hmellor committed 161 days ago

Verified 232ecf2c

[Quantization] fix fbgemm (#42561)

MekkCyber committed 161 days ago

Verified 15b79ea8

make sure the FSDP plugin args are appropriately cast to bools (#42566)

winglian committed 162 days ago

Verified 377a8ee7

[Quantization] per tensor quantization kernel (#42560)

MekkCyber committed 162 days ago

Verified 51c5a7a6

[`CI`] Fix copies (#42571)

vasqu committed 162 days ago

Verified a5c061d2

Allow fallback to loading from Auto"SubProcessor".from_pretrained when model_type can't be inferred from config (#42402)

yonigozlan committed 162 days ago

Verified 675e8763

[SAM3] Compute masks once instead of per-layer, fix fa2 crash (#42543)

yonigozlan committed 162 days ago

Verified 80b408d1

Transformers serve fix (#42570)

LysandreJik committed 162 days ago

Verified 52b988d8

Kernel mapping error resolve (#42466)

Aaraviitkgp committed 162 days ago

Verified 3f174109

fix regression (#42569)

SunMarc committed 162 days ago

Verified 53d2bf6d

Newer Older