Load Instance Fusers Model

Documentation

• Class name: LoadInstanceFusersNode
• Category: instance/loaders
• Output node: False

This node is designed to load and prepare instance fusers models from a specified directory, adjusting their scale according to the provided parameter. It facilitates the dynamic integration of fusers models into the instance diffusion process, enabling enhanced control over the fusion of instance-specific features.

Input types

Required

• model_filename
  • Specifies the filename of the model to be loaded. This parameter is crucial for identifying and retrieving the correct fuser model from the designated directory.
  • Comfy dtype: COMBO[STRING]
  • Python dtype: str
• fusers_scale
  • Determines the scale factor to be applied to the fusers, affecting their influence on the instance fusion process. This allows for fine-tuning the impact of fusers on the generated instances.
  • Comfy dtype: FLOAT
  • Python dtype: float

Output types

• fusers
  • Comfy dtype: FUSERS
  • Provides a list of loaded and scaled fuser models, ready for integration into the instance diffusion process.
  • Python dtype: Dict[str, List[torch.nn.Module]]

Usage tips

• Infra type: GPU
• Common nodes: unknown

Source code

class LoadInstanceFusersNode:
    @classmethod
    def INPUT_TYPES(s):
        return {"required": {
            "model_filename": (get_model_list(constants.INSTANCE_FUSERS_DIR),),
            "fusers_scale": ("FLOAT", {"default": 1.0, "min": 0.0, "max": 10.0, "step": 0.01}),
        }}

    RETURN_TYPES = ("FUSERS",)
    FUNCTION = "load_model"

    CATEGORY = "instance/loaders"

    def load_model(self, model_filename: str, fusers_scale: float):
        checkpoint = load_checkpoint(
            constants.INSTANCE_FUSERS_DIR, model_filename)
        fusers_list = prepare_fusers(checkpoint, fusers_scale)
        fusers = {
            'model_list': fusers_list
        }
        return (fusers,)