invoke-ai / InvokeAI

Invoke is a leading creative engine for Stable Diffusion models, empowering professionals, artists, and enthusiasts to generate and create visual media using the latest AI-driven technologies. The solution offers an industry leading WebUI, and serves as the foundation for multiple commercial products.
https://invoke-ai.github.io/InvokeAI/
Apache License 2.0
23.83k stars 2.45k forks source link

Background mode with reduced GPU usage #148

Closed TingTingin closed 2 years ago

TingTingin commented 2 years ago

Is it possible to add a mode that has reduced GPU usage when inferencing or could someone point me in a direction of the code that would need to change

lstein commented 2 years ago

I don't have an answer to this since my knowledge of CUDA tuning is minimal. Anyone?

lstein commented 2 years ago

If this is a Windows system, perhaps the NVIDIA control panel can be used? https://nvidia.custhelp.com/app/answers/detail/a_id/3130/~/setting-power-management-mode-from-adaptive-to-maximum-performance

TingTingin commented 2 years ago

Main reason I wanted it was so that I could use my pc or record via OBS while rendering I noticed that even having OBS open but idle would slow down renders significantly I think its tied to being at max or close to max vram because I also noticed in the optimizedSD it actually works fine so this could just be merged with https://github.com/lstein/stable-diffusion/issues/62