Henry Cavill?

[lilac kettle]

Can anyone make henry cavills AI voice model? Coz his voice jst too hot and deep like musculine LOL

[candid hamlet]

💡 Tip

You can check if someone already made this model. Try the following:

• Search for it in #1175430844685484042
• Alternatively, in #🔍│find-models:
  • Send " @mossy vortex search (name of the model)", without the ()
  • Use the command /find with @sharp nacelle bot
• Visit https://weights.gg/ and search for the model (login required)

[soft thistle]

@lilac kettle well, if you can prepare a decent enough clean dataset im willing to train it for ya

[lilac kettle]

hiii sorryyy i jst saw your message now

umm how long should the dataset be? Do you need 10-30 minutes of clean voice clips or more?

[lilac kettle]

@soft thistle hi

I have the dataset ready, If you're available to train it for me, I would really appreciate it

[soft thistle]

yes ofcourse send it over in the dm, or just here if you want.

[soft thistle]

yeah nah 10 min is good enough, anything over 30 minute is too much, more than enough for a high quality voice, if the dataset is clean enough the lenght doesnt matter as much

[lilac kettle]

The audio is of 30 mins and the size is 160 mb

Should I trim it down?

[tame steeple]

Actually no, 30 is good too.

The longer, the better as long as the dataset is cleaned properly.

[lilac kettle]

I can't share the file here as it's big

I've uploaded on Dropbox

Does that works?

[soft thistle]

well yeah right, but as you said it too the quality of the idea matters more than lenght after some point ( i would say the 10 minute mark )

[soft thistle]

sure

[lilac kettle]

Here

https://www.dropbox.com/scl/fi/9tx3057nxch689byu8ms5/henrycavillaudio.wav?rlkey=lla0jx1mu6ty1z85mzdxo6ryg&st=7ii827ev&dl=0

[naive crow]

the more the better, but more than 1-2 hours usually gives less noticable improvement given the same quality consistency

[lilac kettle]

umm the question im gna ask is dumb but where can i use rvc v2 models? liek i read that it doesnt work on w okada?

[tame steeple]

You can use V2 models in Applio of course.

But you want to use models with prerecorded audios or on realtime/W-Okada?

[lilac kettle]

ummm realtime

[lilac kettle]

liek when playing games or casual talks

[tame steeple]

Yep, W-Okada is your option for that case.

It supports V2 models

But i recommend using deiteris fork, since it got better performance than OG W-Okada

There you have a guide.

https://rentry.co/ForkVoiceChangerGuide

[lilac kettle]

thanks ill check it out

[tame steeple]

You're welcome bud.

[soft thistle]

@lilac kettle just wanted to give an update, i'll be training the model tomorrow so itll be done in 24 hours max

[lilac kettle]

Sureee no problem and Thankyouu!!!!!

[soft thistle]

https://huggingface.co/kittysec/henrycavill/resolve/main/henrycavill.zip?download=true

@lilac kettle

[lilac kettle]

Tysm!!!!!!!!!!!! It sounds soo goood

I'll try w-okada and let yk!!

[soft thistle]

alr, please do let me know