269 lines
9.2 KiB
Python
269 lines
9.2 KiB
Python
from __future__ import annotations
|
|
|
|
import math
|
|
from typing import Any, Dict, List
|
|
|
|
import comfy.sampler_helpers
|
|
import comfy.samplers
|
|
|
|
|
|
class Guider_AutoGuidanceCFG(comfy.samplers.CFGGuider):
|
|
"""
|
|
AutoGuidance + CFG guider.
|
|
|
|
- "good" model: used for CFG (positive + negative)
|
|
- "bad" model: used as the autoguide reference (positive only)
|
|
|
|
Output matches what ComfyUI guiders return from predict_noise(): "denoised" prediction
|
|
(same convention used by other custom guiders that subclass CFGGuider).
|
|
"""
|
|
|
|
def __init__(self, good_model, bad_model):
|
|
super().__init__(good_model)
|
|
self.bad_model_patcher = bad_model
|
|
self.inner_bad_model = None
|
|
self.bad_conds = None
|
|
self.w_ag: float = 1.0 # paper-style weight; (w_ag - 1) is the delta scale
|
|
|
|
def set_conds(self, positive, negative) -> None:
|
|
# Keep the same conditioning dict style used by CFGGuider-based guiders.
|
|
self.inner_set_conds({"positive": positive, "negative": negative})
|
|
|
|
def set_scales(self, cfg: float, w_ag: float) -> None:
|
|
self.cfg = float(cfg)
|
|
self.w_ag = float(w_ag)
|
|
|
|
def sample(
|
|
self,
|
|
noise,
|
|
latent_image,
|
|
sampler,
|
|
sigmas,
|
|
denoise_mask=None,
|
|
callback=None,
|
|
disable_pbar=False,
|
|
seed=None,
|
|
):
|
|
"""
|
|
Override CFGGuider.sample so BOTH good and bad models go through prepare_sampling()
|
|
and get cleaned up via cleanup_models().
|
|
|
|
This mirrors comfy.samplers.CFGGuider.sample's lifecycle.
|
|
"""
|
|
if sigmas.shape[-1] == 0:
|
|
return latent_image
|
|
|
|
# ComfyUI's prepare_sampling now expects a non-None model_options dict
|
|
# (wrapper plumbing uses model_options.get(...)).
|
|
model_options = getattr(self, "model_options", None)
|
|
if not isinstance(model_options, dict):
|
|
model_options = {"transformer_options": {}}
|
|
self.model_options = model_options
|
|
else:
|
|
# Some setups may have transformer_options missing/None; make it a dict.
|
|
if model_options.get("transformer_options", None) is None:
|
|
model_options["transformer_options"] = {}
|
|
|
|
self.conds = {}
|
|
for k in self.original_conds:
|
|
self.conds[k] = list(map(lambda a: a.copy(), self.original_conds[k]))
|
|
|
|
self.inner_model, self.conds, loaded_good = comfy.sampler_helpers.prepare_sampling(
|
|
self.model_patcher, noise.shape, self.conds, model_options=model_options
|
|
)
|
|
|
|
bad_conds = {}
|
|
for k in self.original_conds:
|
|
bad_conds[k] = list(map(lambda a: a.copy(), self.original_conds[k]))
|
|
|
|
self.inner_bad_model, self.bad_conds, loaded_bad = comfy.sampler_helpers.prepare_sampling(
|
|
self.bad_model_patcher, noise.shape, bad_conds, model_options=model_options
|
|
)
|
|
|
|
self.loaded_models = loaded_good + loaded_bad
|
|
|
|
device = self.model_patcher.load_device
|
|
|
|
if denoise_mask is not None:
|
|
denoise_mask = comfy.sampler_helpers.prepare_mask(denoise_mask, noise.shape, device)
|
|
|
|
noise = noise.to(device)
|
|
latent_image = latent_image.to(device)
|
|
sigmas = sigmas.to(device)
|
|
|
|
try:
|
|
output = self.inner_sample(
|
|
noise, latent_image, device, sampler, sigmas, denoise_mask, callback, disable_pbar, seed
|
|
)
|
|
finally:
|
|
cleanup_conds = {}
|
|
keys = set(self.conds.keys()) | (set(self.bad_conds.keys()) if self.bad_conds else set())
|
|
for k in keys:
|
|
cleanup_conds[k] = (self.conds.get(k, []) or []) + (
|
|
(self.bad_conds.get(k, []) or []) if self.bad_conds else []
|
|
)
|
|
comfy.sampler_helpers.cleanup_models(cleanup_conds, self.loaded_models)
|
|
|
|
del self.inner_model
|
|
del self.inner_bad_model
|
|
del self.conds
|
|
del self.bad_conds
|
|
del self.loaded_models
|
|
|
|
return output
|
|
|
|
def predict_noise(self, x, timestep, model_options: Dict[str, Any] | None = None, seed=None, **kwargs):
|
|
"""
|
|
Mirrors comfy.samplers.sampling_function + cfg_function, but injects
|
|
AutoGuidance delta BEFORE sampler_post_cfg_function hooks.
|
|
"""
|
|
if model_options is None:
|
|
model_options = {}
|
|
# Ensure wrapper-related options always exist
|
|
if model_options.get("transformer_options", None) is None:
|
|
model_options["transformer_options"] = {}
|
|
|
|
# Keep a reference so sample() can reuse it if needed
|
|
self.model_options = model_options
|
|
|
|
positive_cond = self.conds.get("positive", None)
|
|
negative_cond = self.conds.get("negative", None)
|
|
|
|
if math.isclose(self.cfg, 1.0) and model_options.get("disable_cfg1_optimization", False) is False:
|
|
uncond_for_good = None
|
|
else:
|
|
uncond_for_good = negative_cond
|
|
|
|
conds_good: List[Any] = [positive_cond, uncond_for_good]
|
|
out_good = comfy.samplers.calc_cond_batch(
|
|
self.inner_model,
|
|
conds_good,
|
|
x,
|
|
timestep,
|
|
model_options,
|
|
)
|
|
|
|
for fn in model_options.get("sampler_pre_cfg_function", []):
|
|
args = {
|
|
"conds": conds_good,
|
|
"conds_out": out_good,
|
|
"cond_scale": self.cfg,
|
|
"timestep": timestep,
|
|
"input": x,
|
|
"sigma": timestep,
|
|
"model": self.inner_model,
|
|
"model_options": model_options,
|
|
}
|
|
out_good = fn(args)
|
|
|
|
cond_pred = out_good[0]
|
|
uncond_pred = out_good[1]
|
|
|
|
if uncond_for_good is None:
|
|
uncond_pred = cond_pred
|
|
|
|
if "sampler_cfg_function" in model_options:
|
|
args = {
|
|
"cond": x - cond_pred,
|
|
"uncond": x - uncond_pred,
|
|
"cond_scale": self.cfg,
|
|
"timestep": timestep,
|
|
"input": x,
|
|
"sigma": timestep,
|
|
"cond_denoised": cond_pred,
|
|
"uncond_denoised": uncond_pred,
|
|
"model": self.inner_model,
|
|
"model_options": model_options,
|
|
}
|
|
cfg_out = x - model_options["sampler_cfg_function"](args)
|
|
else:
|
|
cfg_out = cond_pred if uncond_for_good is None else (uncond_pred + (cond_pred - uncond_pred) * self.cfg)
|
|
|
|
pos_bad_cond = (self.bad_conds.get("positive", None) if self.bad_conds else None) or positive_cond
|
|
conds_bad: List[Any] = [pos_bad_cond, None]
|
|
out_bad = comfy.samplers.calc_cond_batch(
|
|
self.inner_bad_model,
|
|
conds_bad,
|
|
x,
|
|
timestep,
|
|
model_options,
|
|
)
|
|
for fn in model_options.get("sampler_pre_cfg_function", []):
|
|
args = {
|
|
"conds": conds_bad,
|
|
"conds_out": out_bad,
|
|
"cond_scale": self.cfg,
|
|
"timestep": timestep,
|
|
"input": x,
|
|
"sigma": timestep,
|
|
"model": self.inner_bad_model,
|
|
"model_options": model_options,
|
|
}
|
|
out_bad = fn(args)
|
|
|
|
bad_cond_pred = out_bad[0]
|
|
|
|
ag_delta = (self.w_ag - 1.0) * (cond_pred - bad_cond_pred)
|
|
denoised = cfg_out + ag_delta
|
|
|
|
for fn in model_options.get("sampler_post_cfg_function", []):
|
|
args = {
|
|
"denoised": denoised,
|
|
"cond": positive_cond,
|
|
"uncond": uncond_for_good,
|
|
"model": self.inner_model,
|
|
"uncond_denoised": uncond_pred,
|
|
"cond_denoised": cond_pred,
|
|
"sigma": timestep,
|
|
"model_options": model_options,
|
|
"input": x,
|
|
"cond_scale": self.cfg,
|
|
"bad_model": self.inner_bad_model,
|
|
"bad_cond_denoised": bad_cond_pred,
|
|
"autoguidance_w": self.w_ag,
|
|
}
|
|
denoised = fn(args)
|
|
|
|
return denoised
|
|
|
|
|
|
class AutoGuidanceCFGGuider:
|
|
"""
|
|
Node that constructs the GUIDER object.
|
|
Use with SamplerCustomAdvanced (guider input).
|
|
"""
|
|
|
|
@classmethod
|
|
def INPUT_TYPES(cls):
|
|
return {
|
|
"required": {
|
|
"good_model": ("MODEL",),
|
|
"bad_model": ("MODEL",),
|
|
"positive": ("CONDITIONING",),
|
|
"negative": ("CONDITIONING",),
|
|
"cfg": ("FLOAT", {"default": 5.0, "min": 0.0, "max": 50.0, "step": 0.1, "round": 0.01}),
|
|
# Paper-style parameterization:
|
|
# w_ag = 1.0 => off, 2.0 => moderate, 3.0 => strong
|
|
"w_autoguide": ("FLOAT", {"default": 2.0, "min": 1.0, "max": 5.0, "step": 0.05, "round": 0.01}),
|
|
}
|
|
}
|
|
|
|
RETURN_TYPES = ("GUIDER",)
|
|
FUNCTION = "get_guider"
|
|
CATEGORY = "sampling/guiders"
|
|
|
|
def get_guider(self, good_model, bad_model, positive, negative, cfg: float, w_autoguide: float):
|
|
guider = Guider_AutoGuidanceCFG(good_model, bad_model)
|
|
guider.set_conds(positive, negative)
|
|
guider.set_scales(cfg=cfg, w_ag=w_autoguide)
|
|
return (guider,)
|
|
|
|
|
|
NODE_CLASS_MAPPINGS = {
|
|
"AutoGuidanceCFGGuider": AutoGuidanceCFGGuider,
|
|
}
|
|
|
|
NODE_DISPLAY_NAME_MAPPINGS = {
|
|
"AutoGuidanceCFGGuider": "AutoGuidance CFG Guider (good+bad)",
|
|
}
|