This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! Edit on GitHubThe LaplaceScheduler node generates a sequence of sigma values following a Laplace distribution for use in diffusion sampling. It creates a schedule of noise levels that gradually decrease from a maximum to minimum value, using Laplace distribution parameters to control the progression. This scheduler is commonly used in custom sampling workflows to define the noise schedule for diffusion models.
Inputs
| Parameter | Data Type | Required | Range | Description |
|---|---|---|---|---|
steps | INT | Yes | 1 to 10000 | Number of sampling steps in the schedule (default: 20) |
sigma_max | FLOAT | Yes | 0.0 to 5000.0 | Maximum sigma value at the start of the schedule (default: 14.614642) |
sigma_min | FLOAT | Yes | 0.0 to 5000.0 | Minimum sigma value at the end of the schedule (default: 0.0291675) |
mu | FLOAT | Yes | -10.0 to 10.0 | Mean parameter for the Laplace distribution (default: 0.0) |
beta | FLOAT | Yes | 0.0 to 10.0 | Scale parameter for the Laplace distribution (default: 0.5) |
Outputs
| Output Name | Data Type | Description |
|---|---|---|
SIGMAS | SIGMAS | A sequence of sigma values following a Laplace distribution schedule |