FromIPAdapterPipe (Inspire)

Documentation

• Class name: FromIPAdapterPipe __Inspire
• Category: InspirePack/Util
• Output node: False

The FromIPAdapterPipe node is designed to decompose a previously constructed IP adapter pipeline into its constituent components. This node facilitates the retrieval of individual elements such as the IP adapter, model, and additional features like CLIP vision and InsightFace from a bundled pipeline, enabling further manipulation or analysis of these components.

Input types

Required

• ipadapter_pipe
  • The ipadapter_pipe parameter represents the bundled pipeline from which individual components are to be extracted. It is crucial for enabling the decomposition process.
  • Comfy dtype: IPADAPTER_PIPE
  • Python dtype: tuple

Output types

• ipadapter
  • Comfy dtype: IPADAPTER
  • Extracts the IP adapter component from the bundled pipeline.
  • Python dtype: object
• model
  • Comfy dtype: MODEL
  • Retrieves the model component from the bundled pipeline.
  • Python dtype: object
• clip_vision
  • Comfy dtype: CLIP_VISION
  • Extracts the CLIP vision component, if present, from the bundled pipeline.
  • Python dtype: object
• insight_face
  • Comfy dtype: INSIGHTFACE
  • Retrieves the InsightFace component, if applicable, from the bundled pipeline.
  • Python dtype: object

Usage tips

• Infra type: CPU
• Common nodes: unknown

Source code

class FromIPAdapterPipe:
    @classmethod
    def INPUT_TYPES(s):
        return {
            "required": {
                "ipadapter_pipe": ("IPADAPTER_PIPE", ),
            }
        }

    RETURN_TYPES = ("IPADAPTER", "MODEL", "CLIP_VISION", "INSIGHTFACE")
    RETURN_NAMES = ("ipadapter", "model", "clip_vision", "insight_face")
    FUNCTION = "doit"

    CATEGORY = "InspirePack/Util"

    def doit(self, ipadapter_pipe):
        ipadapter, model, clip_vision, insightface, _ = ipadapter_pipe
        return ipadapter, model, clip_vision, insightface