ImageBatchPermute

Documentation

• Class name: ImageBatchPermute
• Category: image/batch
• Output node: False

The ImageBatchPermute node is designed to reorder a batch of images according to a specified permutation. It allows for dynamic reorganization of image sequences, facilitating operations like shuffling or specific ordering for processing or visualization purposes.

Input types

Required

• images
  • The 'images' parameter represents the batch of images to be permuted. It is crucial for defining the set of images that will undergo reordering based on the permutation pattern.
  • Comfy dtype: IMAGE
  • Python dtype: torch.Tensor
• permute
  • The 'permute' parameter specifies the order in which the images in the batch should be arranged. It directly influences the final arrangement of the images, enabling custom sequences or shuffles.
  • Comfy dtype: STRING
  • Python dtype: str
• start_with_zero
  • The 'start_with_zero' parameter indicates whether the permutation indexing starts at zero. This affects how the permutation pattern is applied to the image batch, aligning with zero-based or one-based indexing conventions.
  • Comfy dtype: BOOLEAN
  • Python dtype: bool

Output types

• image
  • Comfy dtype: IMAGE
  • The output is a reordered batch of images, arranged according to the specified permutation pattern. It enables the dynamic reorganization of image sequences for further processing or visualization.
  • Python dtype: torch.Tensor

Usage tips

• Infra type: GPU
• Common nodes: unknown

Source code

class ImageBatchPermute:
    def __init__(self):
        pass

    @classmethod
    def INPUT_TYPES(cls):
        return {
            "required": {
                "images": ("IMAGE",),
                "permute": ("STRING", {"multiline": False}),
                "start_with_zero": ("BOOLEAN",),
            },
        }

    RETURN_TYPES = ("IMAGE",)
    FUNCTION = "node"
    CATEGORY = "image/batch"

    def node(self, images, permute, start_with_zero):
        order = [int(num) - 1 if not start_with_zero else int(num) for num in re.findall(r'\d+', permute)]
        order = torch.tensor(order)
        order = order.clamp(0, images.shape[0] - 1)

        return (images.index_select(0, order),)