ChatGPT: Renpy + Stable Diffusion + BLIP (2023)

SD webui: Model: BLIP: -Not perfect, as seen from first error (annoying bug that always pops up once, trying to fix - SD actually generates the starting BG) -SD makes new BGs during game (VERY experimental, also only works with developer mode in renpy atm and still seeing if there’s a possible way around this) -BLIP allows character to “see“ (inspired by []): if you send invalid files (non photos) or make character wait too long, character will react accordingly -BLIP2 possible, just ran out of space so can’t test atm lol -Will try with Pyg6b -Will also try Tortoise TTS to integrate voice (will be VERY slow and experimental, can probably search for temporary alternative) [Update: will try ElevenLabs API instead] -It recognized one of the UNLABELED pictures was Australia! (this one is just on ChatGPT,