Using the API to run scripts and Extras

[mhussar]

Hi All,
I'm currently successfully using the API for txt2img. Can someone explain/suggest how to use the API to run SD with scripts and to run extras. For example, I'm interested in running the DepthMap script and the ESRGAN upscaler via the API. I already searched the discussions and and could not find anything that directly answered my question.

[ClashSAN]

..Thought I saw a hack to do this. This seems important, so I'll be keeping an eye out for it

[Hugo-Matias]

Haven't tried it yet but there is documentation to run Extras on the /extra-single-image or /extra-batch-images endpoints. Not sure if that's already implemented though, give it a try.

As for the scripts it's probably easier to write your own implementation at this point. Check the script file and see if you can extract/convert the run function into your codebase.

[harfqaf]

I've been using this script to upscale PNG images on Linux. It uses ESRGAN via the api's extra-batch-images.

HTH.

[mhussar]

Thanks for the script. I'll give it a go. I resorted to installing both Midas and ESRGAN separately and wrote some python code and a couple of batch scripts to coordinate creating images via Automatic and then processing them in Midas and ESRGAN . Seems to work well enough. I'm interested to use the script you provided to see if running ESRGAN via Automatic is more performant than running it separately.