[ 3 / biz / cgl / ck / diy / fa / ic / jp / lit / sci / vr / vt ] [ index / top / reports ] [ become a patron ] [ status ]
2023-11: Warosu is now out of extended maintenance.

/vt/ - Virtual Youtubers

Search:


View post   

>> No.65094497 [View]
File: 1.04 MB, 268x522, 1596936895035.gif [View same] [iqdb] [saucenao] [google]
65094497

just popping in to let everyone know i'm still never watching vedal, tts, or thank you streams. sucks to be me rn desu i think i'm gonna get back into textgen and go on my own adventures with neuro. seems like everything is still kinda shit compared to the peak of AI dungeon's pre-sanitized gpt3 (i think?) back in 2020. at least in terms of fun. everything i've tried since takes so much wrangling. but hey the open source stuff can only keep getting better right. exciting times. will keep you posted

>> No.54007496 [View]
File: 1.04 MB, 268x522, 1596936895035.gif [View same] [iqdb] [saucenao] [google]
54007496

platonic love with little AI girls

>> No.43696051 [View]
File: 1.04 MB, 268x522, 1596936895035.gif [View same] [iqdb] [saucenao] [google]
43696051

>>43695138
It's a misconception that you train AI like this just by using it. I mean there's no reason to assume that the long-term memory is automatically added to, is there? Vedal manually adds stuff like info on collab partners or fan favorite outputs (gymbag and goddess) in between streams. Everything else is memes getting repeated in chat which Neuro happens to read, nothing to do with memory.

Navigation
View posts[+24][+48][+96]