xinyun/ComfyUI
1
0
Fork 0
mirror of https://git.datalinker.icu/comfyanonymous/ComfyUI synced 2026-09-08 00:57:03 +08:00
Code Issues Packages Projects Releases Wiki Activity

Update nodes_optimalsteps.py

Browse Source
This commit is contained in:
ThereforeGames 2025-04-25 18:39:11 -04:00
parent 78992c4b25
commit 7f53c5b0b4
1 changed files with 40 additions and 14 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

54
comfy_extras/nodes_optimalsteps.py
View File

@ -1,9 +1,9 @@
# from https://github.com/bebebe666/OptimalSteps # from https://github.com/bebebe666/OptimalSteps
import numpy as np import numpy as np
import torch import torch
def loglinear_interp(t_steps, num_steps): def loglinear_interp(t_steps, num_steps):
""" """
Performs log-linear interpolation of a given array of decreasing numbers. Performs log-linear interpolation of a given array of decreasing numbers.
@ -18,32 +18,57 @@ def loglinear_interp(t_steps, num_steps):
return interped_ys return interped_ys
NOISE_LEVELS = {"FLUX": [0.9968, 0.9886, 0.9819, 0.975, 0.966, 0.9471, 0.9158, 0.8287, 0.5512, 0.2808, 0.001], NOISE_LEVELS = {"FLUX": [0.9968, 0.9886, 0.9819, 0.975, 0.966, 0.9471, 0.9158, 0.8287, 0.5512, 0.2808, 0.001], "Wan": [1.0, 0.997, 0.995, 0.993, 0.991, 0.989, 0.987, 0.985, 0.98, 0.975, 0.973, 0.968, 0.96, 0.946, 0.927, 0.902, 0.864, 0.776, 0.539, 0.208, 0.001], "SDXL": [12.1769, 9.9182, 7.0887, 4.5944, 2.2473, 0.9020, 0.2872, 0.0738, 0.0197, 0.0020, 0.0]}
"Wan":[1.0, 0.997, 0.995, 0.993, 0.991, 0.989, 0.987, 0.985, 0.98, 0.975, 0.973, 0.968, 0.96, 0.946, 0.927, 0.902, 0.864, 0.776, 0.539, 0.208, 0.001],
}
class OptimalStepsScheduler: class OptimalStepsScheduler:
@classmethod @classmethod
def INPUT_TYPES(s): def INPUT_TYPES(s):
return {"required": return {
{"model_type": (["FLUX", "Wan"], ), "required": {
"steps": ("INT", {"default": 20, "min": 3, "max": 1000}), "model_type": (["FLUX", "Wan", "SDXL"], ),
"denoise": ("FLOAT", {"default": 1.0, "min": 0.0, "max": 1.0, "step": 0.01}), "steps": ("INT", {
} "default": 20,
} "min": 1,
RETURN_TYPES = ("SIGMAS",) "max": 1000
}),
"denoise": ("FLOAT", {
"default": 1.0,
"min": 0.0,
"max": 1.0,
"step": 0.01
}),
},
"optional": {
"custom_sigmas": ("STRING", {
"default": "",
"placeholder": "Comma-separated sigma values"
}),
}
}
RETURN_TYPES = ("SIGMAS", )
CATEGORY = "sampling/custom_sampling/schedulers" CATEGORY = "sampling/custom_sampling/schedulers"
FUNCTION = "get_sigmas" FUNCTION = "get_sigmas"
def get_sigmas(self, model_type, steps, denoise): def get_sigmas(self, model_type, steps, denoise, custom_sigmas=""):
total_steps = steps total_steps = steps
if denoise < 1.0: if denoise < 1.0:
if denoise <= 0.0: if denoise <= 0.0:
return (torch.FloatTensor([]),) return (torch.FloatTensor([]), )
total_steps = round(steps * denoise) total_steps = round(steps * denoise)
sigmas = NOISE_LEVELS[model_type][:] if custom_sigmas:
# Parse the custom_sigmas string into a list of floats
try:
sigmas = [float(s.strip()) for s in custom_sigmas.split(",") if s.strip()]
except ValueError:
raise ValueError("Invalid custom_sigmas format. Ensure it is a comma-separated list of numbers.")
else:
# Use the predefined NOISE_LEVELS
sigmas = NOISE_LEVELS[model_type][:]
if (steps + 1) != len(sigmas): if (steps + 1) != len(sigmas):
sigmas = loglinear_interp(sigmas, steps + 1) sigmas = loglinear_interp(sigmas, steps + 1)
@ -51,6 +76,7 @@ class OptimalStepsScheduler:
sigmas[-1] = 0 sigmas[-1] = 0
return (torch.FloatTensor(sigmas), ) return (torch.FloatTensor(sigmas), )
NODE_CLASS_MAPPINGS = { NODE_CLASS_MAPPINGS = {
"OptimalStepsScheduler": OptimalStepsScheduler, "OptimalStepsScheduler": OptimalStepsScheduler,
} }