SASolverScheduler<T>
Models & Types · AiDotNet.Diffusion.Schedulers
SA-Solver (Stochastic Adams) scheduler using Adams-Bashforth/Moulton methods for SDE sampling.
Inheritance: NoiseSchedulerBase<T> → SASolverScheduler<T>
For Beginners
SA-Solver is a multistep sampler that reuses previous computations (like DPM-Solver) but for stochastic (random) sampling. This gives you the diversity benefits of SDE sampling with the efficiency of multistep methods.
How It Works
SA-Solver applies Adams-Bashforth (predictor) and Adams-Moulton (corrector) multistep methods to stochastic differential equation sampling. This enables efficient SDE sampling with higher-order accuracy using cached function evaluations.
Reference: Xue et al., “SA-Solver: Stochastic Adams Solver for Fast Training of Diffusion Models”, NeurIPS 2023
Constructors
| Constructor | Summary |
|---|---|
SASolverScheduler(SchedulerConfig<T> config, int order) |
Methods
| Method | Summary |
|---|---|
Vector<T> Step(Vector<T> modelOutput, int timestep, Vector<T> sample, T eta, Vector<T> noise) |
|
void SetTimesteps(int inferenceSteps) |