Description
HF: (It's better with text ref)
Comments
How long was the dataset?

46 seconds
It sounds good for 46 seconds it would be cool to see what eleven labs could do
Also did you put it through adobe enhance?

😭 NAH

why is this model actually really good

thanks

its a real thing btw, i resurrected it just for you

im making more memes for gpt-sovits btw

hats off to you

it's actually good

also W for making GPT-SoVITS model

this is nice, I still haven't tried it tho

it says that it will also work with 5 seconds of audio but I can't get the colab working

i actually made this model on colab

its pretty easy to use gpt-sovits tho

i love the person that rated 5

this is literally my dataset


"high quality voice clips"

thats why my model is good

# YOU HAVE A CAR

yeah i let that in the text references too lol

NOOOOOOOOOOOOOOOOOOOOOOOO

guys i think she is not ready to pair
On this did you say car or call?

if you use both, the pronnunciation will be the same

and i used "car"
Wow thats some good accent retention

guys i think its not connected on successfully too 😭😭
How long does it take to make a GPT Sovits model?

you mean on training?

its pretty fast
Yeah
Wow

depending of your dataset

46 seconds of dataset took like 2 minutes

it doesnt take so long for training

i like the accent on "bluetooth"

WAIT

I NEED TO TEST

LAUGH

tf is this

I used the main github and get unknown fatal error for some reason

do you have the colab link you have?


nice..we'll test it out

wait this is the same colab I used yesterday lmao
How many parameters is GPT Sovits?

wdym
Like how many parameters does the pretrained model have

no idea

yes colab hates me

wait its just the uvr weights problem

thats a surprisingly convincing laugh

batch size 7 for sovits and batch size 2 or 4 for gpt
lol
Someone gotta compare Tortoise TTS VS GPT SoVITS
Add a comment
Samples
This model failed processing - generated sample are not available