vllm_omni.diffusion.models ¶

Diffusion model implementations.

Modules:

Name	Description
`anima`
`auk`	AuK diffusion-stage components: the flow transformer, the codec, the pipeline.
`bagel`
`boogu_image`
`cosmos3`
`cosyvoice3_audio`
`diffusers_adapter`	Diffusers backend adapter for vLLM-Omni.
`dmd2`
`dreamzero`
`ernie_image`	ErnieImage diffusion model for vLLM-Omni.
`flux`	FLUX diffusion model components.
`flux2`	Flux2 diffusion model components.
`flux2_klein`	Flux2 klein diffusion model components.
`glm_image`	GLM Image diffusion model components.
`gr00t`
`helios`
`hidream_image`	HiDream Image diffusion model components.
`hidream_o1_image`
`host_weight_contract`	Model-owned declaration for complete final-layout tensor restoration.
`hunyuan_image3`	Hunyuan Image 3 diffusion model components.
`hunyuan_video`	HunyuanVideo-1.5 diffusion model components (T2V and I2V).
`interface`
`internvla_a1`	InternVLA-A1 diffusion model components.
`krea2`	Krea 2 diffusion model components.
`lance`	Lance (ByteDance) diffusion model components.
`lingbot_video`
`lingbot_world`
`longcat_image`
`longcat_video`
`ltx2`
`magi2`	SandAI MAGI-2 Preview diffusion support.
`ming_flash_omni`
`ming_image`	Ming-Image specific diffusion components.
`minimax_h3`	MiniMax H3 audio-video diffusion support.
`mistral_encoder`
`nextstep_1_1`
`omnigen2`
`omnivoice`
`ovis_image`	Ovis Image 7B diffusion model components.
`pi0`	π0 (Pi-Zero) VLA model for vllm-omni.
`pi05`	π0.5 VLA model for vllm-omni.
`pi05_pipeline_config`	Lightweight pipeline topology for π0.5.
`pi0_pipeline_config`	Lightweight pipeline topology for π0 (Pi-Zero).
`progress_bar`	Progress bar mixin for diffusion pipelines.
`qwen_image`	Qwen Image diffusion model components.
`sana_video`
`sana_wm`	SANA-WM diffusion model integration.
`schedulers`
`sd3`	Stable diffusion3 model components.
`sdxl`
`sensenova_u1`
`stable_audio`	Stable Audio Open model support for vLLM-Omni.
`t5_encoder`	Tensor-parallel T5 encoder model.
`utils`
`wan2_2`
`z_image`