Post AYi5KC6GmqtWcGPcQa by theDanielJLewis@noagendasocial.com
 (DIR) More posts by theDanielJLewis@noagendasocial.com
 (DIR) Post #AYi5KC6GmqtWcGPcQa by theDanielJLewis@noagendasocial.com
       2023-08-14T04:21:08Z
       
       0 likes, 0 repeats
       
       @Johncdvorak @adam In the talk about image AIs, have you tried Automatic1111 (https://github.com/AUTOMATIC1111/stable-diffusion-webui/) on your local PC with a specialized model from https://civitai.com/?
       
 (DIR) Post #AaQtpEVHSXjxFHyAU4 by Johncdvorak@noagendasocial.com
       2023-08-14T07:31:02Z
       
       0 likes, 0 repeats
       
       @theDanielJLewis @adam i have not, tell me more
       
 (DIR) Post #AaQtpFUxlD52KapQcy by theDanielJLewis@noagendasocial.com
       2023-10-04T17:06:25Z
       
       0 likes, 0 repeats
       
       @Johncdvorak @adam (Sorry, I just now saw your reply almost two months late!)Automatic1111 is built for Stable Diffusion. It runs totally on your computer and totally offline. It runs best on a computer with a great GPU, or it also runs really well on an Apple Silicon Mac.From a site like civitai.com, you can download a whole variety of models that are built on top of Stable Diffusion and better trained for specific outputs.For example, …
       
 (DIR) Post #AaQu62wxfV1KKsvzIO by theDanielJLewis@noagendasocial.com
       2023-10-04T17:09:28Z
       
       0 likes, 0 repeats
       
       @Johncdvorak @adam … there are multiple models designed specifically to look like Pixar movies or classic Disney animations. There's a model that can make everything look like it's from the movie Tron Legacy. There are models trained for making outputs that look like vector images and would probably convert to vector easily (but all outputs are JPEG or PNG).I've played around with many of these and some can produce some amazing results—loads better than straight DALL-E or Stable Diffusion.
       
 (DIR) Post #AaQuIzx47xESnk9uRU by theDanielJLewis@noagendasocial.com
       2023-10-04T17:11:48Z
       
       0 likes, 0 repeats
       
       @Johncdvorak @adam You can even find things called "LORAs" and "Textual Inversions" that add small details on top of an existing model, and you can use more than one at a time. For example, maybe something specialized in handlebar mustaches, plus something that specializes in reptilian eyes, and something that specializes in making renders of Harrison Ford.