Radiata VS stable-diffusion-webui-rocm

I didn't and don't! Are you saying that can be applied in the a1111 gui? The things I've found by googling it seem to be about a separate UI which uses this optimisation to radically speed up generation.

wow You might be into something here I am going to try this today. Have you tried Lsmith is MIA for 1 month now but is base on TensorRT and it was supper fast when I ran it you need to convert the models to tensorRT format but once they run they are blazingly fast https://github.com/ddPn08/Lsmith

If you just want an accelerated ui, you can check https://github.com/ddPn08/Lsmith/ or https://github.com/VoltaML/voltaML-fast-stable-diffusion which also use the same origina nvidia code. These projects don't do img2img though, you can check in my repo for the img2img pipeline if you need. You need to compile the tensorrt engines for the models first. There are a few steps you can check in their script: export onnx, optimize onnx, compile engine for optimized onnx. I streamlined that a bit and I normally just run my compile.py in docker to build engines.

I just found this amazing GUI that uses the accelerated models of SD to get a speed boost of up to 30%. https://github.com/ddPn08/Lsmith

just released https://github.com/ddPn08/Lsmith

For those that are interested, you can find it at https://github.com/hydrian/stable-diffusion-webui-rocm. Still working on it a bit, but it is pretty stable.