Dw_03

@Viexi Do you still need help? I have a working ROCm enviroment. The only bad thing is that the performances with pytorch in image generation with an rx 9070 (xt)...

@RobertoMaurizzi I suggest you to try a little workaround, if you have time to try something, and use the official docker from AMD. I use it on Fedora (so a...

@kasper93 sorry to bother you. I tried your fork with this commit and the massive memory usage during VAE Decode persists. Also, adding the following flag: `--use-pytorch-cross-attention` increase the memory...

Thanks for the fast reply and the explanations. I forgot to mention that I haven't noticed any problems with using BF16 so far, if it can helps.