Phu Tran
Phu Tran
Replace quantize.py on your machine with this file https://github.com/CompVis/taming-transformers/blob/master/taming/modules/vqvae/quantize.py
@navalguijo I think this issue can be closed
Having this published would also be very helpful for me as well!
Also looking forward to this for accessibility.
> @leejet I've been testing the **SDXL rendering**. I did find some issues: > > 1. for 1024x1024 pictures(or anything above 512), we can see some undecoded latent on the...
Decent images with 4 steps of inference - [Stable Diffusion LCM](https://huggingface.co/SimianLuo/LCM_Dreamshaper_v7) Apache licensed and has a fairly large community. Perhaps a minimal port as an example. - [Tortoise TTS](https://github.com/neonbjb/tortoise-tts) StyleTTS2...
> It could be a good idea to dedicate a space in the README for the GUIs as in llama.cpp. There is one. At the very bottom
Also asserts for all square image dimensions except for 512x512, 768x768, and 1024x1024 ( haven't tested past 1024x1024).
> We still have some problems: > > **PROBLEM 1:** > > Im seeing your code from master and it still use old functions, like this: > > https://github.com/search?q=repo%3Aleejet%2Fstable-diffusion.cpp%20ggml_metal_log_set_callback&type=code >...
Issue goes away when using taesd. Also need to test if the vae fix helps.