huggingface/transformers

Pull Requests Commits

Add GraniteMoeHybrid support for 4.0 (#37658)

Ssukriti committed 221 days ago

Verified 471958b6

[Ready to Merge][HFQuantizer] Squelch pydantic warnings (#37726)

kylesayrs committed 222 days ago

Verified fe29b8c4

Fix incorrect type annotation in get_auxiliary_logits (#37955)

tanuj-rai committed 222 days ago

Verified 46c0e1ff

[generate] Fix `vocab_size` access for multimodal models (#37937)

Jonas committed 222 days ago

Verified d80f53fa

Use T4 single GPU runner with more CPU RAM (#37961)

ydshieh committed 222 days ago

Verified 7819911b

[core] reuse unused reserved cuda memory when loading models (#37920)

gante committed 222 days ago

Verified 3b067a15

More fault tolerant notification service (#37924)

ivarflakstad committed 222 days ago

Verified afbc293e

[D-FINE] Update names (#37957)

NielsRogge committed 222 days ago

Verified 36ca58bf

[docs] logits docstring (#37929)

gante committed 225 days ago

Verified 2932f318

Break weight tying when quantizing input embedding (#37905)

jerryzh168 committed 225 days ago

Verified fa3c3f9c

Aligning modling code for GPT2 to work with vLLM (fallback) (#36934)

ariG23498 committed 225 days ago

Verified 8a0a508f

Add usage example for DINOv2 (#37398)

baldassarreFe committed 226 days ago

Verified e94a4807

🌐 [i18n-KO] Translated `gpu_selection.md` to Korean (#36757)

iambogeumkim committed 226 days ago

Verified d20aa681

Improve performance of `load_state_dict` (#37902)

woct0rdho committed 226 days ago

Verified ee25d57e

[chat] clean code and add base help (#37892)

gante committed 226 days ago

Verified 410aa019

Fix typos in strings and comments (#37910)

co63oc committed 226 days ago

Verified 5b573beb

🚨 rm already deprecated pad_to_max_length arg (#37617)

itazap committed 226 days ago

Verified c80f6526

fixed gemma3 collection path pointing to llama 2 collection. (#37899)

dmgcsilva committed 227 days ago

Verified 7a3e2088

Support `AOPerModuleConfig` and `include_embedding` (#37802)

jerryzh168 committed 227 days ago

Verified 86777b5e

Enhance documentation to explain chat-based few-shot prompting (#37828)

MostHumble committed 227 days ago

Verified c3aeaa80

Fix Qwen3 tp plan with FP8 (#37871)

MekkCyber committed 227 days ago

Verified 36e2e33b

[tests] reset logs in `torch.compile` test (#37894)

gante committed 227 days ago

Verified 8e8025b3

[tests] Test all cache implementations (#37873)

gante committed 227 days ago

Verified 1b222903

Support FlaxPreTrainedModel to load model checkpoint from local subfolder safetensors (#37732)

Melody-coder923 committed 227 days ago

Verified 2c115551

update comment in image_processing_base.py to reference image_process… (#37864)

arjunaskykok committed 227 days ago

Verified 5b223bbc

Fix: reassign in qwen3 moe model (#37848)

linkedlist771 committed 227 days ago

Verified 0dffcb09

uniformize kwargs for VisionTextDualEncoder (#34563)

tibor-reiss committed 227 days ago

Verified 6c5d374d

Fix qwen2-vl-docs. (#37879)

zhanluxianshen committed 227 days ago

Verified 4fc97677

make sure lr is not a tensor (#37881)

winglian committed 227 days ago

Verified 4eb6acc8

fix error for _register_pytree_node in torch2.1.0 and fix bf16 assertion in xpu and npu (#37839)

jiaqiw09 committed 227 days ago

Verified 7be92f9a

Older