viral cedar
Download: https://huggingface.co/V-Mikey-V/Markiplier-V2/resolve/main/Markiplier-V2.zip?download=true
Dataset: 18:47
Sample Rate: 32k
Pretrain: Legacy Core V1.6
Batch size: 8
"Who's coming, huh? If you're coming right now let me see it, let me see it. Show me. HUH. HE, HE WHO IS COMING. You better get out here, you better COME out here. You better come where I am! "
pure ore
holy that really does sound like him
viral cedar
the weird part is is that the dataset i made only uses one of his videos and almost none of the video has yelling at all but it can yell pretty good
the second voice sample doesn't demonstrate it that well though
but in vonovox/realtime it sounds even better
viral cedar
i might try adding his screaming from that video into it and see what happens lol
could be interesting
so far leg V1.6 has been awesome
the quality difference is super noticeable
compared to leg V.15
im actually gonna real quick infer his screaming with this current model and see what happens
pure ore
bet
viral cedar
wow that sounds bad
XD
9 seconds is the best part
yea definitely needs screaming data lol
pure ore
I think adding some screaming will give him the range he needs
maybe some funny noises?
Mark makes funny noises
love this so much
viral cedar
Mark makes funny noises
scenic garnet
Retraining this with:
pure ore
W
