I’ve been putting a lot of thought into automating an AI bot that would do something randomy to generate seed prompts and then feed that through an image generator and post it.
I finished it the other day its called the dallemmybot. I still am getting errors auto-posting to lemmy and pixelfed but the mastodon account updates regularly.
thanks! I’ve just hit my max allowance on chatGPT so that means i need to calm down on the button pressing but I also am looking at getting some used old tesla GPUs to run something locally. I can get a pair for 99 bucks which means that and some hardware to run it should pay for itself in about a year. Would be a fun project to build my own GPU cluster that doesn’t play games :-D
I am working on adjusting the prompting every day so check back often! Im also always looking at adding more model support. Right now I’m using an open source language model to filter search results and that works amazing and I love the idea of using entirely different types of AI together. Right now the script is using 3 for prompt generation.
I’ve been putting a lot of thought into automating an AI bot that would do something randomy to generate seed prompts and then feed that through an image generator and post it.
Go for it! I think Reddit had a bot run community, but back then it was much more random
I finished it the other day its called the dallemmybot. I still am getting errors auto-posting to lemmy and pixelfed but the mastodon account updates regularly.
lemm.ee/u/dallemmybot
https://mastodon.world/@DALLEmmyBot
That looks amazing! Going to share it with friends
thanks! I’ve just hit my max allowance on chatGPT so that means i need to calm down on the button pressing but I also am looking at getting some used old tesla GPUs to run something locally. I can get a pair for 99 bucks which means that and some hardware to run it should pay for itself in about a year. Would be a fun project to build my own GPU cluster that doesn’t play games :-D
I am working on adjusting the prompting every day so check back often! Im also always looking at adding more model support. Right now I’m using an open source language model to filter search results and that works amazing and I love the idea of using entirely different types of AI together. Right now the script is using 3 for prompt generation.
That’s really cool, where do you get used GPUs from?
Also if you have a community or blog set up, I think people would enjoy following along