vllm_omni.diffusion.compile ¶

logger `module-attribute` ¶

logger = init_logger(__name__)

regionally_compile(
    model: Module, *compile_args: Any, **compile_kwargs: Any
) -> Module

Apply regional compilation to a PyTorch model.

Parameters:

Name	Type	Description	Default
`model`	`Module`	The PyTorch model instance to compile	required
`*compile_args`	`Any`	Positional arguments forwarded to torch.compile	`()`
`**compile_kwargs`	`Any`	Keyword arguments forwarded to torch.compile	`{}`

Returns:

Type	Description
`Module`	The same model instance (modified in-place)