Giter Club home page Giter Club logo

Comments (7)

robballantyne avatar robballantyne commented on July 16, 2024

https://stackoverflow.com/questions/73747731/runtimeerror-cuda-out-of-memory-how-can-i-set-max-split-size-mb may be a useful start.

Currently you'll need to set the environment variable on instance startup but I'm pushing an update that will allow supervisord to pick up environment changes without a full restart that will help to debug this.

This being said, it should work fine on a 4090. PyTorch is due an update so that might also help. That will be ready today/tomorrow

from stable-diffusion-webui.

richbrownio avatar richbrownio commented on July 16, 2024

Thanks Rob.

from stable-diffusion-webui.

robballantyne avatar robballantyne commented on July 16, 2024

New builds have been pushed and docs updated for new tags. latest now points to PyTorch 2.2.0 CU11.8.0 but there are also builds for Cuda 12.

Hopefully one of these helps but let me know if not.

from stable-diffusion-webui.

richbrownio avatar richbrownio commented on July 16, 2024

Thanks Rob, it didn't help. I still got the same error message when trying to use an SDXL image resolution of 896x1152. I'm fine if I gen using a lower resolution like 512x786 but with the 24GB 4090 it should have enough memory. I tried to add the 'PYTORCH_CUDA_ALLOC_CONF=max_split_size_mb:512' to the envs and restart but no effect.

from stable-diffusion-webui.

robballantyne avatar robballantyne commented on July 16, 2024

Sorry this hasn't helped Rich, it's late here but I'll have time tomorrow to really dig into this to find out what's going wrong. I'll update here as soon as possible

from stable-diffusion-webui.

robballantyne avatar robballantyne commented on July 16, 2024

Quick update. I've been able to run this today on a 4090 (at vast) with the base sdxl model which was able to generate images at the resolution you specified.

I know that doesn't help much, but I don't believe it to be an issue with the installed software. If you can share your exact settings with models etc I can try again to force a failure.

from stable-diffusion-webui.

richbrownio avatar richbrownio commented on July 16, 2024

Hi Rob - Thanks. Appologies for not getting back to you. I changed the upscaler and have just been creating at a lower resolution for the moment. The next time I jump into the VM at Vast I'll try to replicated the issue and send you the specifics. Thanks for your help.

from stable-diffusion-webui.

Related Issues (20)

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.