Pull Requests huggingface/transformers

[serge] Fix 6 integration tests for model `layoutlmv2` failing with `other` (other (6))

#48808 opened 2026-09-14 22:53 by sergereview[bot]

Fix Glm4MoeIntegrationTest: device_map=auto + MemoryCleanupMixin

#48806 opened 2026-09-14 21:56 by ydshieh

test: run eager/sdpa equivalence on the default masking path for every model

#48804 opened 2026-09-14 19:38 by dltsum

generate: drop an all-ones attention mask from model_kwargs

#48803 opened 2026-09-14 17:21 by SunMarc

Fsdp tp saving

#48802 opened 2026-09-14 16:55 by michaelbenayoun

[Gemma Quant] Preserve Gemma 4 MoE router projection in full precision during quantization

#48801 opened 2026-09-14 16:04 by pculliton

Add regression test for TextToAudioPipeline tokenizer-only models

#48799 opened 2026-09-14 15:46 by Nanduu24

Fix RoPE validation for nested rope_parameters

#48798 opened 2026-09-14 15:36 by hmellor

Fix batched multi-channel to mono conversion in ASR feature extractors

#48795 opened 2026-09-14 14:32 by nandan2003

[apply_chat_template] pass sampling_rate to call

#48794 opened 2026-09-14 13:12 by eustlb

Ep mesh

#48792 opened 2026-09-14 12:47 by 3outeille

Audio processors in processors call

#48790 opened 2026-09-14 09:57 by eustlb

docs: update Simplified Chinese README to match latest English version

#48787 opened 2026-09-14 09:20 by lg114

Fix decoder only path for older bert variants

#48785 opened 2026-09-14 07:53 by Cyrilvallez

Add tensor parallel support for GPT-J

#48783 opened 2026-09-13 23:48 by Siddhant231xyz

Add tensor parallel support for GPT-Neo

#48782 opened 2026-09-13 23:48 by Siddhant231xyz

Add tensor parallel support for OPT

#48781 opened 2026-09-13 23:48 by Siddhant231xyz

Add tensor parallel plan for Nemotron

#48780 opened 2026-09-13 23:48 by Siddhant231xyz

Key the tied-embedding TP rule on the model's input embedding attribute

#48779 opened 2026-09-13 23:14 by Siddhant231xyz

Claude/linear context architecture a35c6h

#48777 opened 2026-09-13 21:14 by Wisk-inc

docs: fix prompt formatting and code alignment in Gemma model doc

#48775 opened 2026-09-13 18:56 by Rattan344

docs: fix grammar typo in ALBERT model card

#48774 opened 2026-09-13 18:12 by Rattan344

fix(wavlm): replace fused F.mha path with explicit submodule calls to…

#48773 opened 2026-09-13 17:18 by p-doshi

[i18n-TR] Translate run_scripts.md to Turkish

#48772 opened 2026-09-13 16:08 by yuefdev

[i18n-TR] Translate model_sharing.md to Turkish

#48771 opened 2026-09-13 15:56 by yuefdev

Fix DogeForCausalLM attending to future tokens under sdpa

#48770 opened 2026-09-13 15:47 by Omarinho12345

[Doge] Fix sdpa attending to future tokens

#48769 opened 2026-09-13 15:19 by dltsum

Add DeepSeek-V4.1 vision support

#48768 opened 2026-09-13 14:38 by malaiwah

docs(zh): add Chinese translation of attention_interface.md

#48767 opened 2026-09-13 14:27 by dltsum

docs(zh): add Chinese translation of cache_explanation.md

#48766 opened 2026-09-13 13:35 by dltsum