Acly

It's likely python version is too old. You can find which one is found in the logs. Also possible you need macOS 14 (I don't have a mac to test...

Yea been waiting for something like that, it's kind of essential.

I did some more tests with this earlier this month, but it feels very restrictive without working img2img. It was in their TODO, but nothing happened for months now...

The automated installer only supports DirectML. That's because I don't have an AMD GPU to test RocM (but I can test DirectML with Nvidia). If you're installing and setting up...

You mean adding a RocM option for the installer? Basically you * add a `ServerBackend.rocm` option [here](https://github.com/Acly/krita-ai-diffusion/blob/main/ai_diffusion/settings.py#L21) (make it linux only) * add a branch [here](https://github.com/Acly/krita-ai-diffusion/blob/main/ai_diffusion/server.py#L189-L192) and put the latest...

Although they don't necessarily match the image dimensions, using layers as inputs has some advantages: * Can easily cut/crop the image to be square and show relevant content * Can...

In v1.18.0 you can now setup IP-Adapter individually for different areas in the image.

Qt (GUI library) number box allows values up to 2^31, so it's actually not simple to fix! Even if it was possible to use the seed value I doubt you...

The seed you see is always computed on CPU. There may be an additional seed for randomness introduced by SDE samplers, which is computed on GPU.

Sounds very similar to #530