towhee.models.action_clip.action_clip.ActionClip

class towhee.models.action_clip.action_clip.ActionClip(clip_model: Module, visual_prompt: Optional[Module] = None, text_prompt_flag: bool = True)[source]

Bases: object

ActionClip model.

Parameters:

  • clip_model (-) – CLIP model

  • visual_prompt (-) – Visual prompt.

  • text_prompt_flat (-) – Flag to control whether to use text prompt.

Methods

encode_text

encode_video

__init__(clip_model: Module, visual_prompt: Optional[Module] = None, text_prompt_flag: bool = True)[source]