Impact Scheduler Adapter

Documentation

• Class name: ImpactSchedulerAdapter
• Category: ImpactPack/Util
• Output node: False

The ImpactSchedulerAdapter node is designed to adapt various scheduling strategies for tasks or processes, allowing for dynamic selection and application of scheduling algorithms based on specific conditions or preferences.

Input types

Required

• scheduler
  • Specifies the primary scheduler to be used, with an option to default to a predefined input scheduler.
  • Comfy dtype: COMBO[STRING]
  • Python dtype: comfy.samplers.KSampler.SCHEDULERS
• ays_scheduler
  • Allows for the selection of an alternative scheduling strategy from a predefined list, including the option to not use an alternative scheduler ('None').
  • Comfy dtype: COMBO[STRING]
  • Python dtype: List[str]

Output types

• scheduler
  • Comfy dtype: COMBO[STRING]
  • Outputs the selected scheduler, which could be the primary scheduler or an alternative one based on the conditions provided.
  • Python dtype: core.SCHEDULERS

Usage tips

• Infra type: CPU
• Common nodes: unknown

Source code

class ImpactSchedulerAdapter:
    @classmethod
    def INPUT_TYPES(s):
        return {"required": {
            "scheduler": (comfy.samplers.KSampler.SCHEDULERS, {"defaultInput": True,}),
            "ays_scheduler": (['None', 'AYS SDXL', 'AYS SD1', 'AYS SVD'],),
        }}

    CATEGORY = "ImpactPack/Util"

    RETURN_TYPES = (core.SCHEDULERS,)
    RETURN_NAMES = ("scheduler",)

    FUNCTION = "doit"

    def doit(self, scheduler, ays_scheduler):
        if ays_scheduler != 'None':
            return (ays_scheduler,)

        return (scheduler,)