Neural nodes use a remote AI service #1987
-
|
I see the ArmorPaint contains some neural nodes. Great work. 👍 But those nodes need a powerful PC. Would it be better if it depended on a remote AI service? |
Beta Was this translation helpful? Give feedback.
Replies: 2 comments 1 reply
-
|
Could add a remote service in the future, but I would not be able to offer it for free, and at the same time I am not a fan of subscriptions. 🙂 Will think about it, but right now the focus is on doing it locally with open-source models so it works even when servers or models are taken down. I want to always include a smaller model like Stable Diffusion 1.5 which takes 4GB of video memory, maybe even lower once some optimization takes place, so it is accessible even on less powerful PCs. |
Beta Was this translation helpful? Give feedback.
-
|
Amazing. |
Beta Was this translation helpful? Give feedback.
-
|
Maybe the remote service could be open source and self hostable. Additionally all the models should also be able to run locally. That way users have a choice of either running locally, self hosting on a more powerfull computer or use some sort of Cloud service where they pay for the usage. |
Beta Was this translation helpful? Give feedback.
Could add a remote service in the future, but I would not be able to offer it for free, and at the same time I am not a fan of subscriptions. 🙂 Will think about it, but right now the focus is on doing it locally with open-source models so it works even when servers or models are taken down.
I want to always include a smaller model like Stable Diffusion 1.5 which takes 4GB of video memory, maybe even lower once some optimization takes place, so it is accessible even on less powerful PCs.