Wagner Bruna

> I think supporting standard json is important for ease of use, but if we could accept keys both with or without quotes, then why not. Sure, one does not...

I'd also suggest moving the top level to `/openai/v1`, to make the target standard more explicit, and avoid conflicts with future alternative standards (e.g. emulating A1111/Forge, as Koboldcpp does). If...

There is a new llama.cpp PR that seems to address this: ggml-org/llama.cpp#15815 > That being said, I don't think there's any sd1.x model that works well at such high resolution...

@zhou7510 , could you try again with `master-309-35843c7` ? It includes many new options for reducing VRAM usage (`--vae-conv-direct`, `--diffusion-conv-direct`, `--offload-to-cpu`). I was able to generate even larger images with...

Working fine on a few tests. Increasing the tile size didn't really help my card, even though it has 16G VRAM, but _lowering_ it to 64 made it almost twice...

Could you please give more details, so I can try to reproduce it here? Which version, model type, and parameters? Are you using a binary release, and on which platform?

I don't have Windows here, but the `sd-master-abb115c-bin-win-avx2-x64.zip` binary seems to work fine on Wine (I don't have the merged model, so I'm using the LCM LoRA): > wine ./sd.exe...

Could you run that command with `-v`, and paste the result here (including the command itself)? Also, what is the last version that do work for you?

So this release is already crashing in that way: https://github.com/leejet/stable-diffusion.cpp/releases/tag/master-0d8b39f ? Please check it carefully, because I can't see how those code changes could cause such a behavior (unless master-0d8b39f...

> No please check the link instead of the number It is master-b1fc16b , isn't it? That's why I asked about the release master-0d8b39f : that's the first release after...