Server execution error: The GPU device does not support Double (Float64) operations!

[ZGainsforth]

Running 1.5 Cinematic Photo plugin on Krita on Snapdragon Elite X, with DirectML on Adreno GPU. The GPU does not seem to support double operations. If you can supply a quantized model it would not only run faster but would run at all. (Using the NPU would be even better but then you have to go all the way to int!) For now, I have to use a CPU only server.

[Acly]

If you can supply a quantized model it would not only run faster but would run at all

What makes you say that? I can't find info about any support for Adreno GPU in any shape or form at all. I don't think it has a chance of working.

The error message isn't related to quantization.