Skip to content

Empty Motion Data

Documentation

  • Class name: EmptyMotionData
  • Category: MotionDiff
  • Output node: False

The EmptyMotionData node is designed to generate a template for motion data with specified frame lengths. It primarily serves the purpose of creating a base structure for motion data that can be further processed or manipulated, providing a standardized format for initializing motion-related tasks within the MotionDiff framework.

Input types

Required

  • frames
    • Specifies the number of frames for the generated motion data. This parameter directly influences the shape of the motion tensor, thereby determining the temporal length of the motion data.
    • Comfy dtype: INT
    • Python dtype: int

Output types

  • motion_data
    • Comfy dtype: MOTION_DATA
    • Outputs a dictionary containing tensors for motion, motion mask, and motion length, establishing a foundational structure for motion data with zeroed motion values and appropriate masking.
    • Python dtype: Dict[str, torch.Tensor]

Usage tips

  • Infra type: CPU
  • Common nodes: unknown

Source code

class EmptyMotionData:
    @classmethod
    def INPUT_TYPES(s):
        return {
            "required": {
                "frames": ("INT", {"default": 196, "min": 1, "max": 196})
            }
        }

    RETURN_TYPES = ("MOTION_DATA", )
    CATEGORY = "MotionDiff"
    FUNCTION = "encode_text"

    def encode_text(self, frames):
        return ({
            'motion': torch.zeros(1, frames, 263),
            'motion_mask': torch.ones(1, frames),
            'motion_length': torch.Tensor([frames]).long(),
        }, )