OptimalSteps scheduler improvements #7801
Conversation
EricBCoding
requested review from
Kosinkadink,
huchenlei,
ltdrdata,
pythongosssss,
robinjhuang,
webfiltered and
yoland68
as code owners
|
Is |
|
Oh, good call. It should probably end with |
Test Evidence CheckIf this PR modifies behavior that requires testing, a test explanation is required. PRs lacking applicable test explanations may not be reviewed until added. Please add test explanations to ensure code quality and prevent regressions. If this PR changes user-facing behavior, visual proof (screen recording or screenshot) is required. PRs without applicable visual documentation may not be reviewed until provided. You can add it by:
|
3 participants
Hi,
I introduced a few quality of life improvements to the OptimalSteps scheduler node from #7584 :
stepsfrom 3 to 1: there are times when you may want to run inference with only 1-2 steps, such as 2-step upscaling @ 50% denoise.SDXLpreset to theNOISE_LEVELSdictionary: the values here are derived from the AYS scheduler and seem to work pretty well.custom_sigmasoptional input: the log-linear interpolation included with this node may prove useful for other models, especially if you can pass in your own sigma values. Supplying a comma-separated string will override the included presets (SDXL, WAN, Flux).