SamplerER_SDE - ComfyUI Built-in Node Documentation

The SamplerER_SDE node provides specialized sampling methods for diffusion models, offering different solver types: ER-SDE, Reverse-time SDE, and ODE. It lets you control the stochastic behavior and the number of computational stages of the sampling process. The node automatically adjusts settings based on the chosen solver type to keep the sampler working correctly.

Inputs

Parameter	Description	Data Type	Required	Range
`solver_type`	The type of solver to use for sampling. Determines the mathematical approach for the diffusion process (default: “ER-SDE”).	COMBO	Yes	”ER-SDE" "Reverse-time SDE" "ODE”
`max_stage`	The maximum number of stages for the sampling process (default: 3). Controls the computational complexity and quality.	INT	Yes	1-3
`eta`	Stochastic strength of SDEs. When eta=0, they reduce to deterministic ODE. Large eta may cause invalid outputs. If this occurs, try decreasing this value. (default: 1.0)	FLOAT	Yes	0.0-10.0 (step: 0.01)
`s_noise`	Noise scaling factor for the sampling process (default: 1.0). Controls the amount of noise applied during sampling.	FLOAT	Yes	0.0-100.0 (step: 0.01)

Parameter Constraints:

When solver_type is set to “ODE” or when eta is 0, the node switches to ODE mode and sets s_noise to 0.0, regardless of the value entered for s_noise.
The eta parameter controls the stochastic strength of both the “ER-SDE” and “Reverse-time SDE” solver types. It has no effect when the solver runs in ODE mode.

Outputs

Output Name	Description	Data Type
`sampler`	A configured sampler object that can be used in the sampling pipeline with the specified solver settings.	SAMPLER

This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! Edit on GitHub

Source fingerprint (SHA-256): 5299ae9b45444cdc7c36bcb3c5e5a0600f9f904e57ae614554033434afdffd30

​Inputs

​Outputs

Inputs

Outputs