search

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
23.61k stars 2.43k forks source link

[enhancement]: Clip Guidance Conditioning #2113

Closed sneccc closed 12 months ago

sneccc commented 1 year ago

Will you guys implement Clip Guidance feature like in dreamstudio?

keturn commented 1 year ago

DreamStudio and some other folks working on similar models have been pushing CLIP guidance pretty hard, so I expect we'll want the option.

There's an example we can look at here: https://github.com/huggingface/diffusers/blob/9be94d9c6659f7a0a804874f445291e3a84d61d4/examples/community/clip_guided_stable_diffusion.py#L113

It does look a lot more resource-hungry than the classifier-free guidance we've been using. Both in computation and memory.

sneccc commented 1 year ago

DreamStudio and some other folks working on similar models have been pushing CLIP guidance pretty hard, so I expect we'll want the option.

There's an example we can look at here: https://github.com/huggingface/diffusers/blob/9be94d9c6659f7a0a804874f445291e3a84d61d4/examples/community/clip_guided_stable_diffusion.py#L113

It does look a lot more resource-hungry than the classifier-free guidance we've been using. Both in computation and memory.

yeah it is resource hungry, but for people that rent gpus A6000,3090 etc it would be fine

github-actions[bot] commented 1 year ago

There has been no activity in this issue for 14 days. If this issue is still being experienced, please reply with an updated confirmation that the issue is still being experienced with the latest release.