[diffusion] model: support MOVA (#17704)

Co-authored-by: gaoyang07 <Gary1546308416AL@gmail.com>
Co-authored-by: cms42 <c@cms42.top>
Co-authored-by: cms42 <44895820+cms42@users.noreply.github.com>
Co-authored-by: Ruixiao Li <cgruixiao@outlook.com>
Co-authored-by: Li Ruixiao(SII) <80368770+Li-dongyang@users.noreply.github.com>

python/sglang/multimodal_gen/runtime/managers/gpu_worker.py

@@ -98,6 +98,9 @@ class GPUWorker:
                 [
                     self.pipeline.get_module("transformer"),
                     self.pipeline.get_module("transformer_2"),
+                    self.pipeline.get_module("video_dit"),
+                    self.pipeline.get_module("video_dit_2"),
+                    self.pipeline.get_module("audio_dit"),
                 ],
             ):
                 if isinstance(dit, OffloadableDiTMixin):
@@ -164,6 +167,8 @@ class GPUWorker:
             if isinstance(result, Req):
                 output_batch = OutputBatch(
                     output=result.output,
+                    audio=getattr(result, "audio", None),
+                    audio_sample_rate=getattr(result, "audio_sample_rate", None),
                     timings=result.timings,
                     trajectory_timesteps=getattr(result, "trajectory_timesteps", None),
                     trajectory_latents=getattr(result, "trajectory_latents", None),