Schedulers play a pivotal role in the diffusion process.
A scheduler is responsible for taking a model’s output and a timestep to return a denoised sample. The timestep is crucial as it indicates where in the diffusion process the step is. Data is generated by iterating forward through timesteps, and inference happens by propagating backward through these timesteps.
They offer a wide variety of results. Depending on your use case, select the one that best fits your needs. For a full list of schedulers accessible via the API please see:
https://docs.scenario.com/docs/schedulers
We generally recommend the default scheduler for a majority of use cases. Adjusting the scheduler is for advanced users.
For a deep dive into Schedulers (also called Samplers) follow this link!
Results Comparison
The following grids of images have been generated using:
- the same public generator (JRPG Adventure)
- the same seed
- the same prompt: "a bear alone in dark woods"
You'll notice than the highest number of steps or guidance don't necessarily give the best results.