#A voice models channel for only HQ models

1 messages · Page 1 of 1 (latest)

shut cedar
#

dont think im trying to be an ass but i think the server would greatly benefit from a channel where only high quality models are posted.
imagine someone trying to find a model of someone and it sounds bad, because the person who made it doesnt have the right experience, wich is fine but i also think that some models needs to be addressed as HQ ones.

for example, if we have 200 kanye models, how can we (or a random user) know if the model is actually good?
the HQ voice models channel will be a guarantee that the hypotetical kanye model is the best one out there.

since theres no real development in RVC as a technology outside of few small performance improvements or quality of life updates, giving the community truly high quality voice models that are PERFECT would be a great advancement.

jolly gardenBOT
#
Vote for this suggestion!
turbid tiger
#

I've suggested the same a while back, a place for Model Masters

shut cedar
sacred trout
#

maybe a vote per model by staff?

shut cedar
#

staffers haves different opinions, that can be done if someone thinks an HQ model is not really HQ

#

but approving them?
would only put a weight on staffers, slow the times the models are out and suffering from bias

glossy brook
shut cedar
glossy brook
#

that's what I heard

shut cedar
#

i dont really know tbh

glossy brook
#

ok

#

what of the Applio devs?

shut cedar
#

applio doesnt change the base code of rvc

glossy brook
#

true

#

so changing the base code is important for further major development?

shut cedar
#

yes

glossy brook
#

ah ok

scarlet summit
#

What do you say?

shut cedar
scarlet summit
#

Also good morning Ila

shut cedar
#

morning to you too :3

#

its because everyone can put it as HQ

shut cedar
#

also, the voice models channel is clogged, new users or generally basically everyone would not care and notice the presence of a tag

orchid reef
#

We'll look into this

untold bone
#

"Please look forward to the pretrained base model of RVCv3, which has larger parameters, more training data, better results, unchanged inference speed, and requires less training data for training." From the readme in the mainline repo