Comic Posted May 22 Share Posted May 22 A lower-powered version of Photoshop's running off of an open source AI image generator with a custom content aware script that runs off of volunteered GPUs. I forgot what it's called but it's when you give a bit of your GPU power to a company that can use it to host their service for everyone else. Quote Link to comment Share on other sites More sharing options...
Tactilis Posted May 22 Share Posted May 22 7 minutes ago, Comic said: A lower-powered version of Photoshop's running off of an open source AI image generator with a custom content aware script that runs off of volunteered GPUs. I forgot what it's called but it's when you give a bit of your GPU power to a company that can use it to host their service for everyone else. @Comic This does not appear to be a question. Or make much sense. 1 Quote Link to comment Share on other sites More sharing options...
BlastOfBN Posted May 23 Share Posted May 23 1 hour ago, Tactilis said: This does not appear to be a question. Or make much sense. I think the OP is asking for a new feature that uses an AI neural network to generate an image in the selected area of a canvas, which is something Photoshop is capable of, since Adobe has been experimenting with AI. @Comic However, if I recall correctly, an AI model would need to connect to the Internet, which is where AI models get their knowledge and information from, to ultimately render an image. Although, you do want it to run off of the GPU’s memory and power, it will probably still need an internet connection to work. I’m not sure if we’re trying to avoid adding features that require an internet connection, though; we’ll have to wait for a response from Rick. Quote Link to comment Share on other sites More sharing options...
Ego Eram Reputo Posted May 23 Share Posted May 23 While you wait, why not render your fill with AI and copy-n-paste into a PDN image? Quote ebook: Mastering Paint.NET | resources: Plugin Index | Stereogram Tut | proud supporter of Codelab plugins: EER's Plugin Pack | Planetoid | StickMan | WhichSymbol+ | Dr Scott's Markup Renderer | CSV Filetype | dwarf horde plugins: Plugin Browser | ShapeMaker Link to comment Share on other sites More sharing options...
Tactilis Posted May 23 Share Posted May 23 8 hours ago, BlastOfBN said: I think the OP is asking for a new feature that uses an AI neural network to generate an image in the selected area of a canvas Yes, I kind of worked that out. I was prompting the OP to think about the wording of their topic title and question. You write a post once but your words are read by 100s or 1000s of people. It helps the readers by writing a clear understandable question; and it helps you because you get more relevant answers. Quote Link to comment Share on other sites More sharing options...
Tactilis Posted May 23 Share Posted May 23 8 hours ago, BlastOfBN said: I’m not sure if we’re trying to avoid adding features that require an internet connection Effects plugins are permitted to make an Internet connection: https://forums.getpaint.net/topic/117240-guidelines-for-publishing-plugins/ Quote Limitations of the Plugin model: ... Effects can open files and read them from your harddrive or Internet URL. Quote Link to comment Share on other sites More sharing options...
BlastOfBN Posted May 24 Share Posted May 24 15 hours ago, Tactilis said: Effects plugins are permitted to make an Internet connection Perhaps we could either make it a built-in feature, or we could have that as a plugin instead. Although, I’m not sure what AI model we would use for it. Quote Link to comment Share on other sites More sharing options...
Rick Brewster Posted May 24 Share Posted May 24 If that's what you want then it sounds like what you need is Photoshop 2 Quote The Paint.NET Blog: https://blog.getpaint.net/ Donations are always appreciated! https://www.getpaint.net/donate.html Link to comment Share on other sites More sharing options...
MJW Posted May 26 Share Posted May 26 On 5/22/2024 at 6:03 PM, BlastOfBN said: However, if I recall correctly, an AI model would need to connect to the Internet, which is where AI models get their knowledge and information from, to ultimately render an image. That's not how it works, at least not for the most widely used AI image generator, Stable Diffusion. All the knowledge and information is contained in files called checkpoints (aka, models), which are typically between about 2GB to 6GB in size. Once the checkpoints are downloaded, no Internet connection is needed to use them. Quote Link to comment Share on other sites More sharing options...
ilbuonmarcio Posted May 31 Share Posted May 31 The other day I was thinking of integrating a local Stable Diffusion (or something similar) with a homemade plugin, but I haven't yet checked if others are already available before jumping in developing one myself. On 5/24/2024 at 2:22 AM, Rick Brewster said: If that's what you want then it sounds like what you need is Photoshop I'd rather extend what's available in Paint.NET than migrating to Photoshop, in my case, for this only feature, the former is already enough for my use case. If it's something we as a community can develop and integrate (without going in conflict with anything, of course), why not? 1 Quote Link to comment Share on other sites More sharing options...
Rick Brewster Posted May 31 Share Posted May 31 6 hours ago, ilbuonmarcio said: If it's something we as a community can develop and integrate (without going in conflict with anything, of course), why not? Sounds good to me! It sounded like you were asking for it to be built-in, which isn't really a possibility Quote The Paint.NET Blog: https://blog.getpaint.net/ Donations are always appreciated! https://www.getpaint.net/donate.html Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.