Nodes/Skoogeer-Noise/KSampler (LoRA Sigma Inverse)
ComfyUI Node

KSampler (LoRA Sigma Inverse)

A ComfyUI node in sampling with 17 inputs and 1 output.

By ttulttul·Created 8 months ago·Updated 2 months ago· 14
KSampler (LoRA Sigma Inverse)
  • model
  • positive
  • negative
  • latent_image
  • LATENT
seed0
steps20
cfg8.0
sampler_name
scheduler
lora_name
min_lora_strength0.00
max_lora_strength1.00
min_lora_step-1
max_lora_step-1
denoise1.00
scale_cfgfalse
clamp_cfg_after-1
Categorysampling

Inputs (17)

NameTypeDefaultDescription
modelMODELThe model used for denoising the input latent.
seedINT00–18446744073709550000The random seed used for creating the noise.
stepsINT201–10000The number of denoising steps.
cfgFLOAT8.00–100Classifier-Free Guidance scale.
sampler_nameCOMBOSampling algorithm.
schedulerCOMBOSigma scheduler.
positiveCONDITIONINGPositive conditioning.
negativeCONDITIONINGNegative conditioning.
latent_imageLATENTThe latent image to denoise.
lora_nameCOMBOLoRA file to load and schedule over sampling.
min_lora_strengthFLOAT0.00-100–100Starting LoRA strength at the beginning of sampling (highest sigma).
max_lora_strengthFLOAT1.00-100–100Target LoRA strength at the end of sampling when sigma reaches its minimum.
min_lora_stepINT-1-1–10000Minimum step index where LoRA is enabled. -1 means no minimum bound.
max_lora_stepINT-1-1–10000Maximum step index where LoRA is enabled. -1 means no maximum bound.
denoiseFLOAT1.000–1How much of the schedule to run.
scale_cfgBOOLEANfalseScale CFG toward 1.0 as LoRA strength increases during active LoRA steps.
clamp_cfg_afterINT-1-1–10000Sets the CFG value to 1.0 after the given sampling step.

Outputs (1)

NameTypeDescription
LATENTLATENT