transformers

583f1942 - Add support for batched Qwen2.5/3-Omni audio generation (#47186)

Add support for batched Qwen2.5/3-Omni audio generation (#47186) * Fix Qwen2.5 Omni * Fix qwen 3 * fix-ci * comments * Update tests/models/qwen2_5_omni/test_modeling_qwen2_5_omni.py Co-authored-by: Eric Bezzam <4757445+ebezzam@users.noreply.github.com> * Add Tests and update Docs * Refactor docs, TTS batch example for qwen3 omni, fix qwen3 omni moe for multi-gpu usage. * Fix Qwen2.5/3 * Fix qwen2.5 * fix-tests-and-atol * fix batch generation * fix batch generation * Fix silent batched audio and wrong talker user-segment in Qwen2.5/3-Omni * fix qwen3 * Update docs/source/en/model_doc/qwen2_5_omni.md Co-authored-by: Eric Bezzam <4757445+ebezzam@users.noreply.github.com> * comments * Ensure backward compatibility for batch size 1 and truncate end of shorter audio for Qwen3 omni. * fix-merge-conflict * Fix qwen2.5 omni after merging with main. And more fix for batch generation. --------- Co-authored-by: Eric Bezzam <4757445+ebezzam@users.noreply.github.com> Co-authored-by: Eric B <ebezzam@gmail.com>