Millennium Whisper · published · build 17517458

Hey everyone,
We've been chipping away at all your feedback as well as working on some of the big content updates that are further on the horizon, but don't be fooled v0.1.1 is a big improvement from v0.1:
We've updated the Large and Standard AI models to new and improved versions. We're happy to say that players with 8GB card who have been using the Standard model will see a HUGE improvements here. Overall characters are less likely to repeat themselves or get confused and are just generally more engaging. This is only the first iteration though and there's many more AI improvements coming very soon!
Speaking of models, you can now choose for yourself which one to use in the new options menu! Before we defaulted to the size your card will be capable of but now the power is in your hands. For instance, you could use the Standard model with a 12GB card and boost the context length instead, allowing the characters a little bit more memory. You can also lower the context length if you're running a lot of simultaneous VRAM intensive tasks (i.e. OBS) and are experiencing crashes.
But that's not all, you can adjust the music levels and text speeds too!
Wondering what the effects of the full moons rampant rumour ruminations are? We found that it was hard to tell how things were changing when full moons took place, so we added a popup that should give you some idea of what's happened.
Let us know what you think and keep whispering!