paddlespeech.resource.resource module
- class paddlespeech.resource.resource.CommonTaskResource(task: str, model_format: str = 'dynamic', **kwargs)[source]
Bases:
object
Methods
get_model_class
(model_name)Dynamic import model class. Args: model_name (str): Model name.
get_versions
(model_tag)List all available versions.
set_task_model
(model_tag[, model_type, ...])Set model tag and version of current task.
- static get_model_class(model_name) List[object] [source]
Dynamic import model class. Args:
model_name (str): Model name.
- Returns:
List[object]: Return a list of model class.
- get_versions(model_tag: str) List[str] [source]
List all available versions.
- Args:
model_tag (str): Model tag.
- Returns:
List[str]: Version list of model.
- set_task_model(model_tag: str, model_type: int = 0, skip_download: bool = False, version: Optional[str] = None)[source]
Set model tag and version of current task.
- Args:
model_tag (str): Model tag. model_type (int): 0 for acoustic model otherwise vocoder in tts task. version (Optional[str], optional): Version of pretrained model. Defaults to None.