NeedsMoar

Well, you're running it on the llvmpipe software renderer which is likely the only thing that provides a high enough vulkan version if the upscaler chose that (and it's not...

The pre-built binary is x64, because using 32 bit windows has been pointless unless your CPU required it since the 64 bit version of Windows 2000 came out. So you're...

What GPU? Intel integrated GPUs didn't have vulkan drivers so that might be the issue, and if you didn't install the AMD or nvidia drivers for whatever real GPU is...

Intel says that processor's integrated GPU has vulkan drivers now, apparently. I'ts going to be unimaginably slow because of the whole 2 processor core thing and trying to shared memory...

What's the output of "pip list -v"? If the packages are somewhere other than appdata\local you're probably installing to another python. What command are you using to install torch? It's...

Glad stuff is working. :-) I don't know why torch is such a pain sometimes but that command above will always do it for me. The tensorflow error appears to...

ah installing wheels is a different command. It's trying to do the build, you want: pip install flash_attn-2.3.2-cp310-cp310-win_amd64.whl in the directory you saved the download

The error you got is because you don't have the CUDA developer's toolkit installed (nvcc is from that) which also needs at least the visual studio community tools installed. You...

You're on a mac with an AMD video card trying to build software that runs on CUDA and needs torch built against CUDA and the NVidia toolkit to build according...

I tried the above a while ago and chainner was producing corrupted ncnn output that neither the newer version of torch the system python had installed, or the old ncnn...