API Error: Blender Plugin calling local Stability Diffusion Server under Linux crashes about ai-render

[benrugg]

Cool, that should help the Automatic1111 people debug it. One potential thought I had just now is that you could try setting a different sampler in the AI Render interface. It's possible that DPM++ 2M is not behaving the same way on Linux. (It's a long shot, but it's something you could try)

[benrugg]

Thanks for posting such a detailed error with the explanation, Windows comparison and the log! Unfortunately this is an error happening in Automatic1111, so I won't be much help. I would suggest posting it on their github.

[Skinless-Glue-Fridge]

I dont know, what exact API post the AI-Render is sending to Stable Diffusion. So what should I tell them ?

[benrugg]

I would just post that same info and those error logs. Hopefully someone will be able to see what the error is just from the logs.

[Skinless-Glue-Fridge]

is there a chance to record or log the call from AI-Render to SD ?

[Skinless-Glue-Fridge]

just opened a ticket at automatic1111@github:
AUTOMATIC1111/stable-diffusion-webui#13235

[benrugg]

Great - I hope they help quickly! Here's what you could do to log what AI Render is sending:

1. Edit the file sd_backends/automatic1111_api.py in AI Render. (You can find AI Render's location on the Blender Add-on Preferences screen).
2. Find the line def do_post(url, data): and add print(data) as the next line (indenting it like the other lines)
3. Launch Blender from the command line so you can see the output. Probably something like this in a terminal window: cd /Applications/Blender.app/Contents/MacOS && ./Blender. (See this for help, if you need it).
4. Render with AI Render
5. View the logged output in the terminal window

[Skinless-Glue-Fridge]

Thanks. I have attached the output of the "print(data)" log of your AI-Render Engine.

Attached you find the call:

output.txt

[Skinless-Glue-Fridge]

Thanks for the idea, but identical result ;-)

[benrugg]

Gotcha