Inference Steps
Learn about using the steps parameter within your creations
Last updated
Was this helpful?
Learn about using the steps parameter within your creations
Last updated
Was this helpful?
The number of steps in the diffusion process determines how many iterations the image goes through.
Each additional step introduces small, incremental changes to the image, allowing for further refinement. You can generate images with up to 150 steps when using on-device processing, providing a high level of control over the final result.
Sogni uses the super efficient DPM Solver scheduler by default, which achieves impressive results in just 20 or 25 steps, ensuring both efficiency and quality. You can select other Schedulers and Time Spacing algorithms in the Advanced tab of the Control Bar.
Tip: To speed up the generation process, begin with a smaller number of steps, typically between 15 and 30. During this stage, you can type or edit your prompt, tweak other settings, and make adjustments. Once you're satisfied with the initial image, lock the seed and increase the number of steps to regenerate the image. This allows you to test if additional steps enhance the detail and quality, helping you refine it to your desired outcome.
You can easily adjust the interface steps by moving the dial or by clicking the arrows within the user interface.
Need help? Ask in our Discord!
For any questions regarding the use of Sogni, tips, suggestions, bugs, sharing art, discussing AI and life, or simply hanging out... Join our Discord community! Join Sogni Discord ✨
