Hacker News new | past | comments | ask | show | jobs | submit login

Others may have reduced the batch size (n_samples) to reduce the memory load. A lower batch size will significantly help with the memory consumption.

This comment: https://news.ycombinator.com/item?id=32710550 talks about running SD with 8GiB of VRAM and mentions needing to reduce this parameter to 1 to get it to output right.




This helped and I finally generated something larger than 256x256 :D thanks


If you're okay waiting a while linger and have plenty of RAM, https://github.com/bes-dev/stable_diffusion.openvino has a somewhat CPU-optimized version as well that relies on system memory rather than VRAM.

My laptop takes about 6 seconds per iteration so it's significantly slower, but if you're willing to wait I bet you'll have a much easier time plugging more RAM into your system than adding VRAM.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: