v2_examples/omni — MoT & cascade models
Runnable examples of the omni models: one request that spans multiple loop types. Two topologies are covered:
- MoT / shared-weight (Cosmos3, BAGEL) — ONE resident instance whose
transformerruns both anar_decodeloop and adiffusion_denoiseloop on the same weights. Every AR token and denoise step is a runtime-visible WorkUnit (the differentiation vs vllm-omni's opaque DIFFUSION stage). - Cascade / separate experts (Qwen-Omni) — three distinct experts on three loop types
(
thinker → talker → vocoder), cross-stage conditioning, streaming codec→waveform.
Run from anywhere:
python3 v2_examples/omni/01_cosmos3.py
| Script | Model | Loops in one request |
|---|---|---|
01_cosmos3.py |
Cosmos3 (shared MoT) | ar_decode (reason) → diffusion_denoise (joint video) |
02_bagel.py |
BAGEL (shared MoT) | ar_decode (generate_text) → diffusion_denoise (generate_image) |
03_qwen_omni.py |
Qwen-Omni (3 separate experts) | ar_decode → ar_decode → audio_decode (text + speech) |
04_interleave_across_loop_types.py |
— | the interleave parity gate holds across AR + diffusion steps |
build_omni_engine() registers all three.