feat: Add support for Qwen 2.5-Omni model bridge - #2296

Closed

martinzhang03 wants to merge 19 commits into

NVIDIA-NeMo:mainfrom

martinzhang03:feat/omni-support

Commits on Mar 17, 2026

[model,docs,test] feat: add qwen3 omni stage1 scaffolding
hbhflw2000
committed
bd917ea
[model,test] feat: add qwen3 omni vision runtime
hbhflw2000
committed
f92d8cc
[model,test] feat: add qwen3 omni audio runtime
hbhflw2000
committed
866b5d5
[model,test] fix: add qwen3 omni functional smoke coverage
hbhflw2000
committed
2b2ef81
[docs,test] fix: align qwen3 omni docs and test paths
hbhflw2000
committed
2ba2be9
[model,test,docs] fix: address qwen3 omni review feedback
hbhflw2000
committed
a8d7ed9

Commits on Mar 18, 2026

[docs] chore: improve qwen3 omni docstring coverage
hbhflw2000
committed
7ae0f55

Commits on Mar 23, 2026

[test,ci] fix: rename qwen3 omni launcher to l0
hbhflw2000
committed
0c1ff02
[test,ci] fix: resolve qwen3 omni workflow conflict
hbhflw2000
committed
6468712

Commits on Mar 26, 2026

[model,test] fix: forward-port qwen3 omni onto upstream qwen layout
hbhflw2000
committed
31e2295

Commits on Apr 8, 2026

Commits on Apr 13, 2026

fix: Qwen2.5-Omni conversion robustness (torchrun/elastic, NCCL, PP=1 collectives)
martinzhang03
committed
8fff03e