Skip to main content
This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! Edit on GitHub
The SamplerSASolver node implements a custom sampling algorithm for diffusion models. It uses a predictor-corrector approach with configurable order settings and stochastic differential equation (SDE) parameters to generate samples from the input model.

Inputs

ParameterData TypeRequiredRangeDescription
modelMODELYes-The diffusion model to use for sampling
etaFLOATYes0.0 - 10.0Controls the step size scaling factor (default: 1.0)
sde_start_percentFLOATYes0.0 - 1.0The starting percentage for SDE sampling (default: 0.2)
sde_end_percentFLOATYes0.0 - 1.0The ending percentage for SDE sampling (default: 0.8)
s_noiseFLOATYes0.0 - 100.0Controls the amount of noise added during sampling (default: 1.0)
predictor_orderINTYes1 - 6The order of the predictor component in the solver (default: 3)
corrector_orderINTYes0 - 6The order of the corrector component in the solver (default: 4)
use_peceBOOLEANYes-Enables or disables the PECE (Predict-Evaluate-Correct-Evaluate) method
simple_order_2BOOLEANYes-Enables or disables simplified second-order calculations

Outputs

Output NameData TypeDescription
samplerSAMPLERA configured sampler object that can be used with diffusion models