ooh, awesome, thx! I’ve finally upgraded my old laptop and will be able to test larger models.
I’ll try to add local image generation later, in about 2–3 weeks.
hello, which model are u using?
the 4o line has incredibly suffocating censorship that I haven’t been able to bypass completely. use the recommended openai models (toggle in settings) - for example, 4.1 or 4-turbo - their censorship is much more 'hacked' (just a bit above vanilla)
I tried 4.1 so far, also their image generation is mesed up lmao.
Can the image generation actually generate new positions or objects ? Or is it just the character being remade ?
