fastvideo.pipelines.pipeline_registry

`fastvideo.pipelines.pipeline_registry`#

Enumeration for different pipeline types.

`get_pipeline_registry`	Get a pipeline registry for the specified mode, pipeline type, and workload type.
`import_pipeline_classes`	Import pipeline classes based on the pipeline type and workload type.

class fastvideo.pipelines.pipeline_registry.PipelineType[source]#

Enumeration for different pipeline types.

Inherits from str to allow string comparison for backward compatibility.

Initialization

Initialize self. See help(type(self)) for accurate signature.

classmethod choices() → list[str][source]#: Get all available choices as strings.

classmethod from_string(value: str) → fastvideo.pipelines.pipeline_registry.PipelineType[source]#: Convert string to PipelineType enum.

fastvideo.pipelines.pipeline_registry.get_pipeline_registry(pipeline_type: fastvideo.pipelines.pipeline_registry.PipelineType | str | None = None) → fastvideo.pipelines.pipeline_registry._PipelineRegistry[source]#

Get a pipeline registry for the specified mode, pipeline type, and workload type.

Parameters:: pipeline_type – Pipeline type to load. If None and mode is provided, will be derived from mode.
Returns:: A pipeline registry instance.

Import pipeline classes based on the pipeline type and workload type.

Parameters:: pipeline_types – The pipeline types to load (basic, preprocessing, training). If None, loads all types.
Returns:: {pipeline_type: {architecture_name: {pipeline_name: pipeline_cls}}} e.g., {“basic”: {“wan”: {“WanPipeline”: WanPipeline}}}
Return type:: A three-level nested dictionary