ironically discussed last night w/ (Certain Nerd);
the issue lies in the fact that both GPT-3.5 and GPT-4's knowledge cuts off in September 2021, whereas stable diffusion itself was released in 2022 and with personal testing the bot has to search to see what it is
so the best we can do solution wise is
make a generic instruction
have the bot provide gelbooru tags (this would only work with waifu diffusion, and that has to be clarified if we do this)
generate the image that way
so as it stands this is actually downright impossible
ironically discussed last night w/ (Certain Nerd);
the issue lies in the fact that both GPT-3.5 and GPT-4's knowledge cuts off in
September 2021, whereas stable diffusion itself was released in 2022 and with personal testing the bot has to search to see what it isso the best we can do solution wise is
so as it stands this is actually downright impossible