Which one is more important, interactivity or story?
>>727764538Woke localization.
>>727764598I need local models specially made for DND shit
>>727764538being an actual videogames with elements such as challenge and a definite end state, which are things that your power fantasy isekai machine almost certainly lacks
Gemini pro cut its free tier and lowered flash from 50 prompts per day to fucking 20. What the fuck am I going to do now?
>>727764538Is this a serious question?
>>727764653I finally invested in a 9070XT and strong CPU. What are the best local models currently for sexy RP?
how do I allow the output to have a variable amount of tokens (and not have it force the set amount)?
I finally cracked the V3.2 filter to do loli shit, with explicit references to age and DEFCON 1 adjectives.This is a problem.
Actually good online models are too expensive to use for long-term storytelling or interactive game needs but they are good for a coom every now and then (yes I am a poorfag)
>>727764843>and not have it force the set amountThat's not a thing I believe, there's only a upper limit on how many tokens can be allocated for a response max.Go to your preset and add something like "Generate responses varying in length from 1 to 3 paragraphs, depending on the flow of the story."
>>727764538censorship to protect minority groups and derail conspiracy theories
>>727764916Please teach me how, senpai.
>>727765047yeah and for some reason my outputs always go for the upper limit. I often have that shit where a sentence gets cut off in the middle, I know there are two or three settings to fix it but those don't work either lol>1 to 3 paragraphsdoesn't work either, and sometimes a single paragraph would be too long as well
>>727764538You should have a balance of both. I still read books with a real narrative when I get tired of power fantasy land.>>727764739You use something other than that dogshit model.
.
