✨│ai-help

[cinder saddle]

I have a very good pc

[proper shale]

Yeah idk what's up

You can probably go into their GitHub and open an Issue:

https://github.com/w-okada/voice-changer/issues

Hopefully devs can see that

And fix it

[wispy lodge]

Looks like some people have experienced the same issue
https://github.com/w-okada/voice-changer/issues/623
https://github.com/w-okada/voice-changer/issues/567

[low shard]

alr, then read https://docs.aihub.wtf/essentials/how-to-make-voice-models/, use the local ways

[wispy lodge]

You could try running as admin probably too

[cinder saddle]

I dont want to make one I want to use a realistic sounding one

[low shard]

My bad, i meant to say https://docs.aihub.wtf/essentials/how-to-make-ai-cover/

whats your gpu btw

[sharp zodiac]

Every time I download rvc, it’s a 7zip file.

It won’t let me run the software.

I have no idea what other download to use.

[cinder saddle]

4080 super

[wispy lodge]

You should click "View information" (or Weitere Informationen) and a button to run will appear

[cinder saddle]

any realistic fun ones you reccomend

[low shard]

ay thats gonna be good

[low shard]

like any funni voices for your ai cover?

[cinder saddle]

girl ones

to troll with ofc

[sharp zodiac]

My covers come out well with a 2060, back when my rvc fuckin worked

[low shard]

so you are asking for ai covers or realtime voice changing?

[cinder saddle]

realtime voice

[low shard]

Ah, then you should use wokada

-rt

[azure marsh]

This interaction has expired, use the command /guides realtime if you wish to see it again.

[low shard]

it's the first one

[cinder saddle]

uhh whats that

[low shard]

another program that allows you to use rvc models in calls in realtime

[cinder saddle]

im so confused so this one isntt that good?

[cinder saddle]

im talking about models btw

ive done eveyrthing that it said on the app

[low shard]

about the girl voice models,
You can search rvc ai voice model:

• #1175430844685484042
• #🔍│find-models
• https://weights.gg/
• https://huggingface.co/models (but watch out cus in hugging face there arent only rvc ai voice models)
• https://applio.org/models
• https://voice-models.com/

but most good realistic girl voices are paid, you should check #1191429836321849435

[cinder saddle]

the paid ones dont even sound that good tho

[brittle wing]

is there a way to make adlibs like this sound less weird when they make like noises n stuff

[craggy brook]

I'm trying to make this sound file on glamrock freddy, but when I adjust it, the sound is distorted with a thin trembling sound. I wonder if you can adjust this sound to suit glamrock freddy, I don't know how to do it.

[spiral bane]

Hi fellas, i don't suppose anyone is able to gimme a link to an arxiv paper on RVC? I've tried looking for the original paper but can't seem to find anything. I've tried training and cloning with the webui already, but would love to read the actual paper and learn more about it

[craggy brook]

I need this

[random falcon]

Please help, why do I hear myself when I connect my Skype? My virtual cable is vb-cable. Please help me. What i am doing wrong?

[rare goblet]

Ayo? @random falcon level 1 !!!

[random falcon]

What do you mean?

Pls people I need help.

[glacial leaf]

hello

is anyone in here that i could help me with some questions ?

[brave swallow]

bro just ask your question don't ask to ask

[brave swallow]

this wrong channel, and if use use voice meter then watch YouTube

[glacial leaf]

so i am completely new to this and ... i was wondering, when I open up say Google chrome, and go to the https://colab.research site...logged in Google... and then hit the download button to begin things.... is anything downloaded to my computer/laptop?

[rare goblet]

Ayo? @glacial leaf level 1 !!!

[brave swallow]

I don't know wait for others to answer

[lament wagon]

Do you have any other simple colabs, similar to Ilaria?

[proper shale]

No, it's dowloaded to the server, since on Colab you use servers

[proper shale]

https://colab.research.google.com/drive/1mHKTGH5e3SAyDSBss1KtiYRbDdQzwSMs - Ilaria RVC Mainline, basically the updated Ila RVC

[lament wagon]

Aight, I'll try it out

[glacial leaf]

so nothing at all is downloaded to my actual computer.?..I was a it worried about that

[proper shale]

Yeah dw about that, everything is stored in there, and all done temporarily

[glacial leaf]

I can't tell you thank you enough.... this has stressed me out all day at work... no joke, i was serisouly considering using a system image to restore on my computerabout to fireup clonezilla and do a

[proper shale]

Oh damn

[brave swallow]

Just for 1 file being downloaded?

(assuming)

chill

[proper shale]

No they just used Colab

[brave swallow]

yep I know

[proper shale]

And thought it was downloading stuff to their computer

[brave swallow]

yeah that's what I'm saying shouldn't be stressful though

[glacial leaf]

i dont mean to sound excessive lol.... just my laptop is just how i like it and i dont wanna mess up my vpn/p2p setup lol

[random falcon]

Youtube is of no use, too many people saying different things, I followed the steps and ended up with nothing

[proper shale]

Yeah that's a valid concern

[brave swallow]

well remains wrong channel go to #🧬│ai-chat tag me there and let me know which software u used

[proper shale]

Can you share what settings you're using on Okada/RVC Realtime's settings and Skype as well?

This will help a ton

[molten dirge]

I was curious on how to reduce noise from my microphone. I also wanna stream with the rvc but I can only play games from my xbox and stream games from my xbox. I can use the rvc on my laptop but I was curious on how to stream me gaming and use the rvc on my laptop. Never streamed before so any tips for that would be good, thank you

[brave swallow]

to livestream I assume you need OBS, for mic I use Nvidia broadcast

[proper shale]

Noise: Basically, get SteelSeries Sonar or, if you have an NVIDIA GPU, NVIDIA Broadcast is pretty decent. Okada's built in noise suppression also works fine, I suppose, but if it isn't enough get the others.

Streaming: I believe you'd need a capture card to get the XBOX footage to the PC (?) and then put it all on OBS to stream

[brave swallow]

does okada provide virtual mic?

to use as input audio in OBS

[proper shale]

no, you need to download a VAC as well (virtual audio cable)

[brave swallow]

[proven hill]

or even better: https://huggingface.co/spaces/TheStinger/Ilaria_RVC

[proper shale]

I forgor

[proven hill]

peak performance dont forget the bribe i gave you

[proper shale]

It should be working fine as is, don't see anything wrong here

[proper shale]

WHYYY ar eyou trying to incriminate me here 😭

[proven hill]

https://imgur.com/t9cxTNd

[half trellis]

hi I have a problem when I want to add my custom ai model already made by rvcDisconnected by zip it doesn't show up in my model selection tab has anyone had a similar problem?

[proper shale]

the true appearance of ilaria (shes a cat)

[proper shale]

What are you using to do conversion

[half trellis]

of what?

[proven hill]

real

[half trellis]

im wanna send a video but i cant send here

[rare goblet]

Ayo? @half trellis level 1 !!!

[proper shale]

Now you can

[proper shale]

ah, you're using locally

you need to upload the files inside the .zip, not the zip itself

[half trellis]

i have files in zip

[proven hill]

that program is old

[proven hill]

you need the .pth and added .index

only them

[quaint dock]

Rayman???

[half trellis]

okey but where is the problem?

[proven hill]

i told you

[proven hill]

yes

[quaint dock]

Sorry for texting here, I only watched captian laserhawk for rayman

[proven hill]

same lol

[odd shale]

Same

Tho i did play Rayman Origins on my old ps vita.

[lilac dawn]

Can someone provide me with a training RVC?

[brave swallow]

where is the problem/question?

[lilac dawn]

-_-

better?

[brave swallow]

the question isn't clear

you need software or looking for guidance?

you have a dataset and want someone to train it or what

[low shard]

idk if it bugged then

[peak tusk]

I already got an answer, I have to send an email

[proven hill]

i disassociate myself from this conversation, i have no interest in talking and engaging in a conflict, mikus is acting on his way and not talking for me.
so please avoid conflict first, and do not think i am involved in this bullshit.

thanks.

[low shard]

oh hell nah world war 3 😭

[brittle wing]

can someone help me with rvc

each time i clone a voice it sounds horrible

https://tenor.com/view/facepalm-really-stressed-mad-angry-gif-16109475

[rare goblet]

Ayo? @brittle wing level 1 !!!

[lilac dawn]

i just want to train a model on a website

[brittle wing]

https://streamable.com/p5kbov

[brave swallow]

you need for Collab

@proven hill does ilaria rvc provides api ? I want to use it to generate TTS for my livestreams

[proven hill]

nope

[brave swallow]

[rare goblet]

Ayo? @brave swallow level 7 !!!

[brave swallow]

in future?

[proven hill]

i dont plan on doing it tbh.
maybe Applio has an API system? idk i didnt follow much new development.

[brave swallow]

yes they do

but I don't know if it's online hosted

ill recheck

[proven hill]

its probably locally hosted

[brave swallow]

that's the issue

Livestreams already encode using GPU

so having another project open will blew the pc

[proven hill]

for tts i dont suggest rvc tho

you should try SoVitsGPT

[brave swallow]

I wanted to inference my voice

[brave swallow]

ok thanks for suggesting I'll look into it

[proven hill]

best tts option

[brave swallow]

can I use my model on it

or train over

😂

[proven hill]

or if youre looking for a simpler solution you can try ElevenLabs wich offers API and simply voice cloning.
But its paid

[proven hill]

you need to re-train with the SoVits technology, much harder than RVC but better result

[brave swallow]

well I'm circling to not pay

[proven hill]

1 minute of data is fine but it needs to be transcribed

[proven hill]

then SoVitsGPT

[brave swallow]

I see , sounds linke tokenizer

[proven hill]

should be much lighter on GPU too

[brave swallow]

oh cool why Rvc more popular

[proven hill]

btw, arent you the person in the HuggingFace server or am i wrong?

[brave swallow]

I assume yes but i don't think we met before

[proven hill]

RVC is simpler to train on and to infer, more models and especially voicechanger options.
The popularity is because of that, sadly, because this means we are stuck with a very old tech

[proven hill]

nope, i saw your username there tho

[brave swallow]

yeah , im kinda active guy

I manage couple communities

[proven hill]

always AI?

[brave swallow]

no Ai was my first attempt this month

last time I learned about ai was 3 years ago, I learned the concept only how neural works
AGI VS ANI
SUPER AI

but I'm learning so fast I got also my my voice trained 👀

[low shard]

are you tryna use tts in calls?

[proven hill]

i think streams

[brave swallow]

rate it of 10

[brave swallow]

nope chat TTS

in my livestream

not live inference like #🔍│help-w-okada

[low shard]

ohh

then idk lol

[brave swallow]

thry can type !tts then say what they want in my voice

and everyone in the livestream hear it

[brave swallow]

no worries I'll code it😂

[crude flame]

You gonna need a good gpu if you want gptsovits, a game or something and a stream all running at the same time

[brittle wing]

can someone help me with rvc each time i use it it sounds very bad

[rare goblet]

Ayo? @crude flame level 19 !!!

[brave swallow]

hmm, yeah that's the issue, Idk why can't rely on cpu as it's not learning?

[proven hill]

its not that resource intensive the inference

[crude flame]

Really? ive never messed with it before so idk

[brave swallow]

there's not inference using cpu? so I put the project in vps

[proven hill]

me too but i documented a bit.
italian is not supported so my interest is very low

[proven hill]

good question, you should probably ask the main (and only) developer

(only iirc)

[brave swallow]

I'll check with applio first

if they do offer it before going and starting over

they have api so probably works on Linux

cpu

only

[proven hill]

let me know, im curious about it

[brave swallow]

applio docs

heh

there's hope

[proven hill]

4GB of VRam seems pretty pretty low...

[brave swallow]

vram means GPU isn't??

how much cost 4GB vram rent?

[proven hill]

VRAM is the Virtual RAM of the GPU

[proven hill]

pretty cheap, i dont think they even do it

[brave swallow]

who are they??

I never hosted before

[proven hill]

mhhhhhhhhhh

paperspace maybe?

[rotund prawn]

hi

[brave swallow]

https://docs.applio.org/get-started/alternatives/

[brave swallow]

hi

[rotund prawn]

Is there a way to download easyGUI locally? like applio and Ilaria

[low shard]

isnt applio api only for the models searching

[brave swallow]

they have internal api , and cli fork
for local hosted one

[brave swallow]

they saying yes, but slow, no config needed it's automatically switched to cpu if there's no gpu detected

[proven hill]

whats wrong with mine

[rotund prawn]

no no

you are the best

[brave swallow]

good question

+1

[rotund prawn]

❤️

[proven hill]

anyway yes, iirc there was a way but i dont remember, you better check their github

[low shard]

i thought u were talking abt https://applio.org/api/app/docs

my bad

[proven hill]

regarding this: yes its default rvc but cpu is slow

[brave swallow]

no this one for the domain they have atm already

[rotund prawn]

okey thanks

[proven hill]

no prob!

[vernal radish]

also do yk what this error is

Input #0, mp3, from 'C:\Users\THESHA~1\AppData\Local\Temp\0617lozo9kzn.MP3':
  Metadata:
    encoder         : LAME3.101 (beta 2)
  Duration: 00:01:06.14, start: 0.000000, bitrate: 128 kb/s
    Stream #0:0: Audio: mp3, 44100 Hz, stereo, fltp, 128 kb/s
[NULL @ 0000023ab63387c0] Unable to find a suitable output format for 'C:\Users\The'
C:\Users\The: Invalid argument
2024-06-17 17:04:20 | INFO | infer.modules.uvr5.modules | Executed torch.cuda.empty_cache()\```

[rare goblet]

Ayo? @vernal radish level 1 !!!

[brave swallow]

Logical question:
I've found hosts that pay per hour

does taking max specs and train you model worth it?

in terms of time and cost maybe costs 10$ per H but you get the model trained with 1000 epoch maybe 🤔

[proper shale]

Again, training locally is much better

[proper shale]

seems like something related to the audio you're using... try something else, or re-export in Audacity perhaps

[brave swallow]

depends if electricity so expensive * time spent 👀

++++ if you hurry to see results 😂 this damage me today

[thorny spindle]

[Voice Changer] Waiting generate pipeline...
[Voice Changer] Pipeline is not initialized.
[Voice Changer] Waiting generate pipeline...
[Voice Changer] Pipeline is not initialized.
[Voice Changer] Waiting generate pipeline...
[Voice Changer] Pipeline is not initialized.

Please help

[proper shale]

delete pretrain and model_dir folders, restart okada

should install the things it needs, most likely screwed up on initial installation

[proven hill]

RIGHT

that was the potential fix

i couldnt scratch it off my mind

[thorny spindle]

it didint work why?

[rare goblet]

Ayo? @thorny spindle level 2 !!!

[proper shale]

if it skips the installs, id just use https://rentry.co/RVCRealtimeGuide atp

[proud elbow]

what? you cant divide a single training workload into different systems, though multi gpu setup is possible.

you might consider renting one or few A100 from AWS, vast.ai, etc. it is much faster and has overkill 80 gb vram so that u could train several models at the same time.

[brave swallow]

I just checked that A100 its like 3$ per hour nice

yeah my question was, if it's sounds more like crypto training so probably training on 5 different places would be faster to get results

unless it's based on last point then it make sense why won't make difference

[proper shale]

nah

[proud elbow]

it's still much different, crypto mining doesnt cost much labor other than setting up the rig, unlike this case

[brave swallow]

I meant the logic

if its random try and fail
then 5 people/ machine try to find the number will be faster than having one

[proud elbow]

it doesnt apply on ur case at all

[brave swallow]

well I figured out this 😄

[rapid umbra]

hi i have a bit of a problem when i try oppening the Realtime vc it dosnt open the gui , any fixes? i have tried uninstalling and then installing it again

[proud elbow]

the software development and testing (SDLC) is more accurate analogue, where you invest some time and costs on it until acceptance testing to satisfy the what the client needs

[brave swallow]

you mean how G and D works yep

[last crystal]

Hey so I'm like extremely new to RVC

I'm trying to use these voices for TTS in something called SillyTavern

But when I download these, there's no audio files

just a pth file and an index file

I have no clue what to do here

[brave swallow]

this what you need

but is there a tts button or it just ask to upload audio?

[last crystal]

it'd help if I could upload images so I could show you what I see but yeah to upload new voices you need to upload the audio files

[brave swallow]

show me

probably u got rvc only as it sounds

[last crystal]

I can't upload images since I don't have perms but I got an ElevenLabs API key

[rare goblet]

Ayo? @last crystal level 1 !!!

[brave swallow]

hmm

[last crystal]

just realized i should probably blur the API key that I have lol

[brave swallow]

ok

[last crystal]

there we go

Maybe I have to go into the files and drag it in somewhere IDK

[brave swallow]

sometimes yes, but is there an option called models?

[brave swallow]

btw just upload the files , give it name and try it first

[last crystal]

Can I upload the pth and index files into elevenlabs?

[brave swallow]

I don't know about it, I thought you asked about " sillyTavern"

let me see

[last crystal]

cause i'm using an API key from ElevenLabs

And I can see all my ElevenLabs voices in the model list

[brave swallow]

this for TTS

you need it

you will put the index and pth in the project

I assume in the voice section

[last crystal]

need what for TTS

ElevenLabs?

[brave swallow]

then the project will use elevenlabs to generate TTS then inference the voice of the guy

[brave swallow]

yep

[last crystal]

where is the project

[brave swallow]

the one u use?

[azure marsh]

Documentation

AI HUB Docs

https://docs.aihub.wtf/

🍏 Applio Docs

https://docs.applio.org/

✨ More guides

• Full AI Voice Model Training Guide (local) by Christopher Villanueva
• Model training with Mainline RVC by Raven (ravencutie21)
• AICoverGen Colab Guide by Eddy (Spanish Helper)
• Create a model with RVC disconnected (colab) by Angetyde

[midnight bison]

In case you don't know who's singing/rapping whose part:

1. Johnny Cage - Future (Intro, Choruses, ad-libs, Outro)
2. Franklin Clinton - Ace Hood (Intro, Verse 1)
3. Chase - Meek Mill (Verse 2)
4. Jax Briggs - Rick Ross (Verse 3)
5. Snoopy - T.I. (Verse 4)
6. Inkling Boy - Wiz Khalifa (Verse 5)
7. King Dedede - Birdman (Verse 6)
8. Tails - French Montana (Verse 7)
9. Darwin - 2 Chainz (Verse 8)
10. Riley Freeman - Lil Wayne (Verse 9)
11. Huey Freeman - Boo (Verse 10)
12. Carl Johnson - Ace Hood (Verse 11)
13. Donkey Kong - DJ Khaled (Intro, ad-libs)

[crude flame]

Yo can we not post stuff with instrumentals

[midnight bison]

Could you at least let me know your thoughts on this AI cover?

[crude flame]

It sounds

[tame mica]

It sounds

[midnight bison]

It sounds what?

You gonna let me know what you think or no?

[midnight bison]

Well? What are your thoughts?

[crude flame]

it has sound, there are words being said, there are multiple verses, it has at least two instruments

[midnight bison]

It's called an AI cover. It covers songs.

At least let me know what part you enjoyed (or not enjoyed), please?

[crude flame]

i dont like rap in general so

[vernal radish]

what are the best settings for these?

[gilded charm]

I just fixed it, it seems like every time I select japanese_hubert_base it bugs the entire model, even if I change it back

[rare goblet]

Ayo? @gilded charm level 4 !!!

[gilded charm]

Alright!

[tribal vigil]

Why can't I fusion model in Ilaria Zero rvc

[rare goblet]

Ayo? @tribal vigil level 1 !!!

[proper shale]

Maybe the two models you're trying to merge were made in different sample rates? That would explain it

[hidden dew]

Should i keep running it or should i sto

p?

[lilac dawn]

where can i find the train.log on Ilaria RVC?
https://35472d4100deda6dd7.gradio.live

or can someone provide me with a begginer friendly RVC training ONLINE software?

[sand osprey]

#1159949278270193734

[lilac dawn]

what is the ideal lenght of an audio file for training?

[brave swallow]

I assume thig g/total, show us kl and mel

[brave swallow]

10 min

[molten dirge]

is w okada better then rvc then for noise supression built in?

[proper shale]

If you want that noise suppression, yeah

[lilac dawn]

what do i do now

[marsh grove]

guys download isnt woprking

can someone help

https://colab.research.google.com/github/iahispano/applio/blob/master/assets/Applio.ipynb#scrollTo=-7cQtXouqpQi

this is the one i use and when i try to doiwnload the inferred audio, it says File not available on this site

[lilac dawn]

which one do i download?

[brittle wing]

question: i forgot to save the G and D files on my last training session.. but i have the latest epoch file.. can i still continue training?

[tame mica]

you havent finished your training yet

retrain for 1 more Epoch to get your finished model

[tame mica]

i don't think so

[brittle wing]

shit

whats the best training colab atm ?

[lilac dawn]

i tried and it crashed again

[lilac dawn]

can someone hop on vc and help me train a model and post it on hugging face please?

[brave swallow]

nope

[proper shale]

I'd say either RVC Disconnected or Applio work fine

Just gotta limit your training sessions, while using RVC Disconnected

[proper shale]

that's the training files, normally the epochs r saved in the weights folder

[spare solstice]

guys wheres the hugging face version of rvc

[lean bane]

what should my gpu settings be on applio if i have 32gb ram? or should i not use it at all

depends on the voice, IE shadow the hedgehog/ Haru okumura sound out of breath regardless unless I use different audio files

[proper shale]

https://huggingface.co/spaces/TheStinger/Ilaria_RVC

[proper shale]

keep them as they are, i suppose

if you're referring to batch size, the thing u should care about is your GPU's VRAM

[spare solstice]

thank you mj

[proper shale]

yw :)

[lean bane]

ok, thanks

[proper shale]

once you know your VRAM, you know like, how much batch size you can go up to, sometimes even a bit more

[low shard]

@proper shale how to make rvc v4

[digital ingot]

[proper shale]

train a 3 second pretrain and call it the world's most versatile rvc pretrain

ez

[low shard]

ok but can it e gril the voice then?

[proud elbow]

minecraft cat meow model

[low shard]

lmfao u reminded me that guy who made the spongebob pretrain

[proper shale]

yes e gir

[low shard]

ngl i feel like there is already a model like that, i remember seeing a minecraft door model

[dense glacier]

hello

where can i get some girl voices?

[tacit badger]

wtf???

  File "multiprocessing\process.py", line 315, in _bootstrap
  File "multiprocessing\process.py", line 108, in run
  File "D:\ilariarvc\infer\modules\train\train.py", line 234, in run
    torch.load(hps.pretrainG, map_location="cpu")["model"]
  File "D:\ilariarvc\runtime\lib\site-packages\torch\serialization.py", line 791, in load
    with _open_file_like(f, 'rb') as opened_file:
  File "D:\ilariarvc\runtime\lib\site-packages\torch\serialization.py", line 271, in _open_file_like
    return _open_file(name_or_buffer, mode)
  File "D:\ilariarvc\runtime\lib\site-packages\torch\serialization.py", line 252, in __init__
    super().__init__(open(name, mode))
FileNotFoundError: [Errno 2] No such file or directory: 'assets/pretrained_v2\\'```

I think I installed broken Python version or broken pip update

when I click train model I get error

[low shard]

You can search rvc ai voice model:

• #1175430844685484042
• #🔍│find-models
• https://weights.gg/
• https://huggingface.co/models (but watch out cus in hugging face there arent only rvc ai voice models)
• https://applio.org/models
• https://voice-models.com/

if there isnt one, you can:

• #1159289738314919936
• #1191429836321849435
• make it yourself with our docs guides https://docs.aihub.wtf/

Most good ones are paid btw

[obtuse rapids]

can i make voice model but not use in google colab?

[proper shale]

yeah, you can use in https://huggingface.co/spaces/TheStinger/Ilaria_RVC for example

[obtuse rapids]

what do i go on

[brittle wing]

is it faster or slower than colab

[rare goblet]

Ayo? @brittle wing level 10 !!!

[obtuse rapids]

hello?

[proper shale]

model loader, to load models

[brittle wing]

i think the guy wants to MAKE models

[proper shale]

definitely faster, since it uses the A100

[obtuse rapids]

yes

[brittle wing]

.

[proper shale]

ah

[brittle wing]

can i train on huggingface btw?

[proper shale]

[proper shale]

yeah not a good idea 😭

[obtuse rapids]

awww

[brittle wing]

ive been using rvc disconnected and its a pain in the ass really

ah

im currently having an issue with it

[proper shale]

like, you can train locally if you have a good GPU

[brittle wing]

i would but i have a 960 😭

its an old old pc

from like 2015 lmfao

works fine for me, since im not into games that require lots of processing power

[proper shale]

why does your zip folder have a D: folder

[brittle wing]

i dont know

ive just zipped and uploaded it

to my drive

its loaded in

[proper shale]

it uhm

shouldn't have a D: folder

[brittle wing]

i have no idea why its done that

[proper shale]

i think that just screws it up

[proper shale]

id say try zipping again

[brittle wing]

it might be because im using some free software from the microsoft store to unzip and zip stuff lmao 😭

ill just try the standard windows function

oh btw @obtuse rapids if you want to make models

use RVC disconnected

[obtuse rapids]

i ues winRAR

[rare goblet]

Ayo? @obtuse rapids level 3 !!!

[brittle wing]

there are various tutorials you can find in the server

[proper shale]

just use 7zip or WinRAR

[brittle wing]

yeah it works now but i still have to restart my runtime 😭

since the destination path exists already or something

[obtuse rapids]

NOT google colab

[brittle wing]

oh mb

[obtuse rapids]

it ok

[frosty pawn]

soemone help me pls bro my voice changer trippin

[violet heron]

!howtoask

[patent trellis]

How To Troubleshoot

__**GIVE CONTEXT.**__ 📝

• Don't simply mention your issue, like "my rvc is not working".
• Describe the step you are on, what you're trying to do, the RVC you're using, a screenshot, etc.
• The more context, the better.

__**BE POLITE.**__ <:matsuripray:1159685390156967936>

• Don't be desperate. You can ping a Helper, but if they ignore, they aren't available/don't know the answer.
• It's okay if you're frustrated, but don't take it into this server.
• Don't DM without prior consent.

__**BE PRODUCTIVE.**__ 🤝

• Don't ask for every little instruction. Put your own effort & test things by yourself.
• Don't ask to ask.
• Check if your answer is a Google search away/on our guides website.

If you've changed your mind about your message, feel free to delete/edit it.

[frosty pawn]

my issue is its trippin

its good tho i fixed it

[brittle wing]

??

i disconnected and deleted my runtime before this

why does that already exist

is it the fact its a multispeaker model when it isnt?

how do i fix that

fixed it by manually editing the code

made the else statement the same as the elif above

because the else statement was causing it to become a multispeaker model

yeah nvm this is so broken for me

[obtuse rapids]

can i make voice with this? https://github.com/RVC-Project/Retrieval-based-Voice-Conversion-WebUI

[brittle wing]

beyler zip oluşturuyorum fakat

eklenmiyor

rvcde

[rare goblet]

Ayo? @midnight bison level 3 !!!

[magic steeple]

i have some problems can somebody help me

😦

[violet heron]

Say what you need help with then

[midnight bison]

@magic steeple you'll tell me what you thought of the U.O.E.N.O. cover, right?

[foggy idol]

which .pth file am i supposed to copy? i have pth in the logs and also a ton of g and d pth in the logs/project

[violet heron]

the one that has your model name and just pth

So like if I made a model titled RickAstley I would download the “RickAstley.pth” model

[inner cloak]

Keep getting "HTTPError Traceback (most recent call last)
<ipython-input-4-eed3d0216092> in <cell line: 44>()
44 for url in pretrained_urls:
45 filename = os.path.join(output_directory, os.path.basename(url))
---> 46 urllib.request.urlretrieve(url, filename)

6 frames
/usr/lib/python3.10/urllib/request.py in http_error_default(self, req, fp, code, msg, hdrs)
641 class HTTPDefaultErrorHandler(BaseHandler):
642 def http_error_default(self, req, fp, code, msg, hdrs):
--> 643 raise HTTPError(req.full_url, code, msg, hdrs, fp)
644
645 class HTTPRedirectHandler(BaseHandler):

HTTPError: HTTP Error 401: Unauthorized"

for custom pretrained download idk what to do

For anyone wondering, it has to do with:

Failed to download /content/program_ml/rvc/pretraineds/pretraineds_custom/G_KLMv7s_Batch2_32k.pth: HTTP Error 401: Unauthorized
Failed to download /content/program_ml/rvc/pretraineds/pretraineds_custom/D_KLMv7s_Batch2_32k.pth: HTTP Error 401: Unauthorized
Failed to download /content/program_ml/rvc/pretraineds/pretraineds_custom/G_KLMv7s_Batch2_40k.pth: HTTP Error 401: Unauthorized
Failed to download /content/program_ml/rvc/pretraineds/pretraineds_custom/D_KLMv7s_Batch2_40k.pth: HTTP Error 401: Unauthorized
Failed to download /content/program_ml/rvc/pretraineds/pretraineds_custom/G_KLMv7s_Batch2_48k.pth: HTTP Error 401: Unauthorized
Failed to download /content/program_ml/rvc/pretraineds/pretraineds_custom/D_KLMv7s_Batch2_48k.pth: HTTP Error 401: Unauthorized

so doesn't matter tbh

I just use titan tbh. original cell in colab was erroring out and not downloading anything.

[lilac nest]

help please when i say somthing it does like (exmpl thing:hello im name) but it goes like he-... i-.... na-.... me-...

[gusty kestrel]

@brittle wing, I have found 9 results that match your search!

rigosaenz

Jeffy (SML)

• Uploaded: <t:1706570335:d>
• Server: AI HUB
• Likes: 0
• Lang: EN

Epochs

1000

Technology

RVC

Algorithm

Rmvpe

Requested by fuckkissah · Jun 18, 2024, 10:48 PM

[lilac nest]

also it says "Frequent errors occur. Please check if the model of the framework being targeted is loaded"

[gusty kestrel]

@brittle wing, I have found 7 results that match your search!

pdepepsi

Mordecai

• Uploaded: <t:1692662400:d>
• Server: AI Hispano
• Likes: 0
• Lang: ES

Epochs

750

Technology

RVC

Algorithm

Rmvpe

Requested by fuckkissah · Jun 18, 2024, 11:30 PM

An error occurred

Sorry, I could not find models that match your search "pops"

[brittle wing]

how can i tell which model in search is gonna be the best one

is it just which ever has the most epochs

[gusty kestrel]

@brittle wing, I have found 7 results that match your search!

k3jim

Mordecai

• Uploaded: <t:1698099434:d>
• Server: AI HUB Brasil
• Likes: 0
• Lang: PT

Epochs

350

Technology

RVC

Algorithm

Rmvpe

Requested by fuckkissah · Jun 18, 2024, 11:38 PM

[brazen merlin]

Someone help me. Can I get the tools to make an ai voice on mac?

[pearl grove]

hey can someone help me with an error

[inner cloak]

Idk not sure. I removed klmv in the code and it's fine. Might just be me?

[pearl grove]

why is this an error: creds = Credentials.from_service_account_file(config_path, scopes=scope)

[low shard]

Are you using google colab? If so, send the colab link

[rare goblet]

Ayo? @pearl grove level 1 !!!

[pearl grove]

yeah I am but im in a call and they said I should install applio

[low shard]

You prolly used an outdated google colab, oh if someone is already helping u and u have a good pc go for it

[pearl grove]

thanks

[proven hill]

or you can use this in your browser without installations: https://huggingface.co/spaces/TheStinger/Ilaria_RVC

[jaunty mason]

wtf? error

[proven hill]

thats strange

what model are you using?

[torpid siren]

Hi guys i have a question

[sacred spindle]

i have an answer

[torpid siren]

I have trained a voice model and want to test it in some text to speach website or make a cover but dont know where i can do that. Can i get some help with that?

[torpid siren]

😄

[sacred spindle]

if you have a download link to the pth file you can use https://huggingface.co/spaces/TheStinger/Ilaria_RVC

[torpid siren]

I dont have download link i have pth file on my desktop. Can it work with that? Sry if question is stupid im new to this

[sacred spindle]

yes a .pth file works

[torpid siren]

https://prnt.sc/vkblkpS-G2Ab
What i do wrong? Im not sure what index file is

[midnight bison]

In case U.O.E.N.O (pun intended) who's singing/rapping whose part:

1. Darwin - Future (Choruses, Verse 6)
2. Snowball - Rocko (Verse 1)
3. Patrick - Rick Ross (Verse 2)
4. Rocky - A$AP Rocky (Verse 3) (makes sense since Rocky and A$AP Rocky both have 'Rocky' in their name)
5. Charmy - Wiz Khalifa (Verse 4)
6. Jax - 2 Chainz (Verse 5)

[torpid siren]

Nvm i think i got this

Awesome work

[midnight bison]

What parts did you enjoy? What parts did you not enjoy?

[torpid siren]

First 5 verses rocking by my opinion. Im not biggest fan of last one.

[midnight bison]

And why is that?

[torpid siren]

Preferences i think. For me this voice is idk too robotic

[midnight bison]

In case you don't know who's singing/rapping whose part or which part:

1. Donkey Kong - DJ Khaled (Intro)
2. Sonic - T-Pain (Choruses)
3. Shadow - Trick Daddy (Verse 1)
4. Roy (from Pokemon Horizons) - Rick Ross (Verse 2)
5. Ryuji/Skull - Plies (Verse 3)
6. Dark Pit - Young Jeezy (Verse 4)
7. Incineroar - Ludacris (Verse 5)
8. Fox - Busta Rhymes (Verse 6)
9. Linus - Big Boi (Verse 7)
10. Darwin - Lil Wayne (Verse 8)
11. Gumball - Fat Joe (Verse 9)
12. Falco - Birdman (Verse 10)
13. Johnny Cage - Rick Ross (Verse 11)

[torpid siren]

ok thanks. Can i ask how long you train your models?

[rare goblet]

Ayo? @torpid siren level 2 !!!

[empty epoch]

I use RVC with nvidia broadcast to cancel the background noise and I get BSoD/freeze occasioanlly pls help

[open stag]

Colab

[rare goblet]

Ayo? @open stag level 1 !!!

[austere horizon]

hi

[brittle wing]

is there a way i can remove the who do i trust

or will ti sound weird

[fading drum]

anyone have strategies to speed up training without getting a new and more powerful gpu? using rvc1006 nvidia

[proud elbow]

shorten the dataset length and/or increase batch size but needs more vram, also too high values would yield worse generalization (read some GAN article for more explanation)

[robust meadow]

When everyone uses DeEcho-DeReverb in UVR, what window size and aggression settings do you set?

[fading drum]

and batch size goes up in increments of? is it 1:1 or 1:0.5 or something else?

bad thing to use ratios, if i set it to 6 would that be 6gb or vram?

[bleak yoke]

does this look right?

[low shard]

@midnight bison POST your ai cover on yt and share the YT LINK ONLY in #1159290752195633273

[brittle wing]

default settings usually work fine tbh

[spice thorn]

@crude bolt i fixed voicechanger

[rare goblet]

Ayo? @spice thorn level 4 !!!

[obtuse rapids]

can i make it so one voice go in one mp3 and one go it a mp3

[viscid raven]

does anyone know if there is a way to get an english voice model, and let's say change it's accent so that it flawlessly speaks another language?

[glacial pollen]

There is no perfect solution to that.
Closest you can do is to not use the index file

or simply set it to 0 ( the index ratio slider )

It will try to mimick whatever you feed the model to the best of it's abilities

[viscid raven]

ok thanks for the help

[glacial pollen]

✌️

[glacial pollen]

To straighten it up, it is not so that high values worsen the generalization, it is the not adequate for your dataset values do hinder the generalization and promote divergence

Too little batch will provide " too little information " for the hifigan to work out a proper solution and might cause divergence or being stuck.
Too little batch = more noise in gradients

Too high batch will cause " too much of information " so, again, model might overfit too fast to the data as it's exposed to too much of it and in a rather quick manner, in the same time, not being able to learn all that's needed in time.

Best way to visualize it is, imagine a student learning;
You want to learn just enough per given time frame to not overstrain

[jade portal]

when i try and select a model (even the default ones) it says protobuf parsing failed and then shows an error screen

[glacial pollen]

First of all, share what rvc release or fork you are using.
2nd thing is, post a screenshot of console ( log )

[jade portal]

im using MMVCServerSIO_win_onnxdirectML-cuda_v.1.5.3.18a.zip and i cant post a screenshot of my console log

[glacial pollen]

Well, this is not rvc ( retrieval based voice conversio )

It is W-okada you're talking about

#🔍│help-w-okada
is where you wanna go

[jade portal]

shit

oopsies

[glacial pollen]

However

[jade portal]

yuh

[rare goblet]

Ayo? @jade portal level 1 !!!

[glacial pollen]

Protobuf is one of the key components

means either something's outdated or missing

What gpu you have? @jade portal

cause directml is meant for amd

whereas Cuda

for Nvidia

[jade portal]

rtx 3080

[glacial pollen]

In that case, wait, imma link you the right version

[jade portal]

alrigh thanks

[glacial pollen]

https://huggingface.co/wok000/vcclient000/resolve/main/MMVCServerSIO_win_onnxgpu-cuda_v.1.5.3.18a.zip?download=true

@jade portal there

[jade portal]

thanks man

[glacial pollen]

for future, Cuda and DirectML are libraries / platforms for acceleration and calculation, in AI etc. ( but not only )
DirectML ( when cuda is also an option ) is rather for cpu / or amd gpus

so if you see Cuda, that's what you want for nvidia

[jade portal]

ohhhhhh ok

[devout niche]

hello, i seem to have a problem with setup or something and i dont know how to fix it

when i try to open .bat for this, it gives out an error " running with the system python. Nie moPress any key to continue" if i press anything it coses and if i wait nothing happens

also setup.bat immidietly closes itself when i try to open it

[crude bolt]

oh this looks much better ? what did you do 😮

[glacial pollen]

Polish user I see

Can help you out with it. but in a short, you don't have all what's needed to run rvc

most importantly, python ( or either it's not added to the path

Either way, gonna leave it here as I don't have time for green-idling users.
Install python 3.10.6 or add it to the path in case you already have it. ( Just google " how to add python to the path " )

[jaunty mason]

Childish Gambino (RVC) 1k Epoch ?

too much storage or not working

too

why

llaria rvc x rvc2 v2 error not fkn working

[odd shale]

Umm, i suggest you using Ilaria RVC Zero or Applio Colab.

You can find Ilaria RVC Zero space on #📰│dev-updates channel

[low shard]

i think the 1st pic is of Ilaria RVC Zero

[sand blaze]

why I can't find the names I'm looking for in the model section

[odd shale]

I've realized that guy is using a model that got spaces on the name.

[odd shale]

Because if you're looking for a model of something but it doesn't appear, it means there's no model like that.

[low shard]

I analyzed the model files, seems like it has some useless things, you should try to:

1. Download it
2. Unzip it
3. Delete every file except the pth (UltimateFrank18270.pth which is the voice file), and the added.index (added_IVF1379_Flat_nprobe_8.index which is the accent of the voice)
4. Zip those only 2 files, and upload it to huggingface giving credits to the original creator
5. After you uploaded it, right click the download icon next to the .zip file and do copy link adress
6. Paste it in Ilaria RVC HF Space

This SHOULD work

[low shard]

that isn't the issue, as you can use them in ilaria rvc zero, also i tried myself without spaces

[sand blaze]

I'm doing it by watching a video right now, and when the man in the video writes turkhish in the model section, a lot of Turkish names appear, but when I type turkhish, nothing appears.

[rare goblet]

Ayo? @sand blaze level 1 !!!

[odd shale]

Outdated video.

[low shard]

A video? Most Youtube Videos are outdated

[odd shale]

🐢👍

[low shard]

You can search rvc ai voice model:

• #1175430844685484042
• #🔍│find-models
• https://weights.gg/
• https://huggingface.co/models (but watch out cus in hugging face there arent only rvc ai voice models)
• https://applio.org/models
• https://voice-models.com/

if there isnt one, you can:

• #1159289738314919936
• #1191429836321849435
• make it yourself with our docs guides https://docs.aihub.wtf/

[sand blaze]

The video is a bit old but why were the models deleted?

[low shard]

The server got deleted once because of copyright issues, might be that some models got lost, anyways try to use all the sites i told you

you may find it

[sand blaze]

thanks for helping bro

[weak vale]

Hey, I am trying to download the RVC mainline model, I just tried to run the go-web.bat file and got an error in the terminal stating - "C:\Windows\system32>runtime\python.exe infer-web.py --pycmd runtime\python.exe --port 7897
The system cannot find the path specified." Do I need to set up python on my pc to run this?

[devout niche]

I REINSTALLED AND EVERYTHING WORKED OUT, THANKS ANYWAY

[rare goblet]

Ayo? @devout niche level 1 !!!

[devout niche]

caps sry

[glacial pollen]

alr

[winter seal]

happened while training with uka pretrain

[proud elbow]

have u entered correct pretrain G & D files, and the matching sample rate?

[winter seal]

i put te correct pretrain files but i dont know what sample rate i should use

[proud elbow]

look the pretrain file names (e.g. containing 32k)

[winter seal]

o

ok

[empty epoch]

hi how do i delete rvc? im trying to download the CUDA version

[weak vale]

Hey, I am trying to download the RVC mainline model, I just tried to run the go-web.bat file and got an error in the terminal stating - "C:\Windows\system32>runtime\python.exe infer-web.py --pycmd runtime\python.exe --port 7897
The system cannot find the path specified." Do I need to set up python on my pc to run this?

can anyone help me with this please? I tried to look it up but couldnt find anything ^

[violet heron]

RVC should have a version of python that came with it

[brave swallow]

i think you mean pytorch?

[weak vale]

@violet heron do i have to install python locally before running the go-web.bat file?

also thx btw

[violet heron]

it should of came with RVC

Did you unzip everything

[weak vale]

@violet heron yeah I did... that is super weird

[rare goblet]

Ayo? @weak vale level 1 !!!

[weak vale]

https://huggingface.co/lj1995/VoiceConversionWebUI/blob/main/RVC1006Nvidia.7z i installed from this link, this is the one your talking of right?

[violet heron]

Yes

Try installing it from the GitHub isneyad

[weak vale]

damn I really want to use this for vocals for my music but even running it as admin wont work

[weak vale]

you have a direct link for that? appreciate it

[violet heron]

Wait nvm that links the same one 💀

[weak vale]

fuuuuuuck 😢

[violet heron]

Try unzipping it again and running

[weak vale]

kk will do

@violet heron i'm using 7zip to unzip the rar file, any chance that could cause issues during unzipping?

[violet heron]

also make sure it’s not stored on OneDrive

[violet heron]

7zip should work

[weak vale]

@violet heron changing location worked, thank you so much! havent run it yet but at least the GUI popped up... I dont have an NVIDIA, but I have a 7800xt, it says its running on cpu instead of gpu, will that cause any issues?

[violet heron]

cpu inference is stupid slow

But is usable

You can’t train tho

[weak vale]

RIP, I figured it would be slow as hell. thanks man

[violet heron]

If you want a faster option I would go with Ilarias HF space is easier and inference takes like 20 secs

-spaces

[azure marsh]

<:huggingface:1179800228946268270> Hugginface Spaces

• 🆕 Ilaria RVC Zero, by thestingerx Huggingface Spaces
• 🆕 RVC⚡ZERO, by r3gm Huggingface Spaces
• Applio, by IA Hispano Huggingface Spaces
• AICoverGen, by r3gm Huggingface Spaces
• Advanced RVC Inference, by r3gm Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces

More commands: -audio, -colabs, -kaggle, -local, -overtraining, -realtime, -rvc, -help

[weak vale]

AttributeError: 'NoneType' object has no attribute 'dtype' when trying to convert audio, anyone run into this?

[lyric blaze]

-paperspace

[azure marsh]

RVC v2 Tutorial for Paperspace

💡 Click on the button below to open the guide.

Credits: LollenApe

[brave swallow]

is there a why to see how is the XTTS finetune going while training ?

[rain prism]

so my friend blue @gusty elm runs the start_http.bat file but it keeps crashing anyone knows how to help?

[rare goblet]

Ayo? @rain prism level 5 !!!

[pastel oak]

https://cdn.discordapp.com/attachments/1159290161683767298/1159607580184821760/issue.mp4?ex=6673fa77&is=6672a8f7&hm=2fab39a7d43b745b0c82f1a8912724d38da78c0fa28f5f45f1528756bc491e31&

do this

[brittle wing]

guys how to dowload voices

[rain prism]

question do i need python?

@pastel oak

[pastel oak]

if you downlaoded wokada it already comes with python installed

rvc prepackage aswell

[somber viper]

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• RVC Mainline, by Hina Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• EasyGUI, by rejects Google Colab
• UVR5 NO UI for Google Colab, by Eddy Google Colab

More commands: -audio, -hf, -local, -kaggle, -overtraining, -realtime, -rvc, -help

[rain prism]

we didnt download wokada or rvc prepackage

[drowsy swan]

where is that one document that helped get beginners set up

[plush compass]

What settings do you all use for DeEcho Dereverb? I'm using 15 for 272, what do others think?

[alpine lotus]

i'm a good qc

[vague sandal]

i have an issue where rvc will somehow take the voices of the people im vc in with

and output it making it basically dupe what they just said

[alpine lotus]

make sure to use your mic as input and the virtual cable as output.

[brittle wing]

if i have 2 hours of speech will random coughs and stutters affect a model if

i train a model on it

[alpine lotus]

bc that might happen if you have virtual cable output as input and virtual cable input as output.

[brittle wing]

they say that i am a good qc too but im not a model master yet xd

and im a minor to do commissions and i dont want to lol

[thin stump]

Coughing is bad for any dataset and stuttering should be considered speech like "I-I-I-I I think you are..." so you can leave it in

and we also have the deplosive tool with RX11. Adjust the frequency range if needed

[brittle wing]

is there a way to like automatically remove noises like that

cuz theres too much audio

😢

idont wana go thru it all

[thin stump]

I go through my datasets carefully and yeah unfortunately it takes up time

[brittle wing]

is there a way to import all the labels into audacity in 1 track then

cuz rheres alot of like

blank space

on the one i made before this i manually removed all the gaps and it took wayyy too long

and this is like 3x longer

[thin stump]

I have a 1 hour of dataset and it's split into 44-ish files so I'm slowly reviewing those. You can remove blank spaces if you just use audio labels on audacity. Noise gating helps get rid of some of the low noises

[brittle wing]

no like i wanna remove them so it can listen to all of them without the gaps inbetween

just to save time

[brittle wing]

this ones split into like 600 lol

thats why i wanna

just listen to them all back to back

so i kno which parts to remove

[thin stump]

Oh I see. Nope, I didn't use audacity labels yet so I was the one who chose to split it into 1-2 minute clips each

600 is a lot to go through because it's like .10-.15 seconds

[brittle wing]

i deleted all the files that were under 3 seconds

its still alot tho

like 400

[glacial pollen]

careful with noise gating

it can damage subtle tiny bits at certain parts

same for labels, it's not supervised so it can sometimes break natural pauses

[brittle wing]

all of it is around the same volume and i put a limiter on it

can u not import all audacity files into 1 track why do they all go into seperate tracks

[thin stump]

I do it manually drag and drop

[brittle wing]

yeah but theres too many files to manually do it in my case

[thin stump]

You should still have the files before the labels. Split it for yourself so it's more manageable

[brittle wing]

It looks like this

theres alot of gaps

i dont think theres a way to quickly skip over them

[thin stump]

I would say eq that but I can't recommend that on audacity

[brittle wing]

i have

fl

fl studio

[thin stump]

friend request

[brave swallow]

could someone tell me why its not working !!!!!! , its ubuntu

applio

0 progress

100% cpu

[thin stump]

Yeah they can use either truncate or use labels. He didn't follow the aihub guide so he wasn't aware there was a threshold he can adjust for db

default settings was -20db for truncate

[thin stump]

I did read this before so I do hear what you're suggesting btw #✨│ai-help message

[brittle wing]

help, it keeps saying connection errored out

[rare goblet]

Ayo? @brittle wing level 4 !!!

[proven hill]

ilaria rvc can include space in the name of the models now

[exotic cedar]

What could I do to improve the vocal range of a model for voice changer purposes?

Like are there any merging tricks or tips etc

[warm mulch]

guyss

which setting is better?

the fo method

[hot lagoon]

need help

[proper shale]

get ffmpeg if you haven't already

[proper shale]

I mean... merging seems to screw up the model's range, so... I think the best thing you can do is train the model yourself with a large variety of pitch ranges, if possible ofc.

[glacial pollen]

There is one thing you could attempt but I can't promise or guarantee anything.

Provided your model is perfectly made to 100% reflect the dataset, and I am serious about it, perfectly made, then try to render some stuff which are on the verge of model's pitch limits

add a bit of that to your already existing dataset and re-train

• the training itself, if not done correctly ( say, without batch_size range finding ) will have the pitch range hindered anyway

[slate compass]

my audio works perfectly

but not in discord vcs

anyone help me pls

[exotic cedar]

Thanks for the tips @glacial pollen and @proper shale

[glacial pollen]

✨

[proper shale]

You're welcome :)

[proper shale]

Maybe use a different audio cable

[jovial tulip]

Hii, I'm trying to configure AICoverGen (https://github.com/SociallyIneptWeeb/AICoverGen?tab=readme-ov-file) but this problem appears for me:

running build_ext
building 'fairseq.libbleu' extension
error: Microsoft Visual C++ 14.0 or greater is required. Get it with "Microsoft C++ Build Tools": https://visualstudio.microsoft.com/visual-cpp-build-tools/
[end of output]

note: This error originates from a subprocess, and is likely not a problem with pip.
ERROR: Failed building wheel for fairseq
Failed to build fairseq
ERROR: Could not build wheels for fairseq, which is required to install pyproject.toml-based projects

I already downloaded Visual Studio and this problem still occurs

[granite falcon]

Gusys i have a problem i have made model using applio and been trying to test it but everytime i try to download it it says" its not a zip" or something like that, ive already compresed the dataset on zip and uploaded it to drive , copied the link and made it available to use to anyone with the link but still error pops out

Is drive no longer accepted or what?

I want to upload it to hugginface if thats the case but i dont know how to use it or upload it there

[ancient field]

i have a problem aswell

[glacial pollen]

's what you need

it's in a tab where you pick individual ones rather than full packages etc

You gotta pick the msvc ver 140 essentially

Other than that, win 10 sdk and basic functions of C++ compilation tools or however it is localized in your lang

[violet heron]

Maybe instead of saying you have a problem maybe state your problem

[granite falcon]

Anyone?

[ancient field]

wait web smthin

[rare goblet]

Ayo? @ancient field level 1 !!!

[ancient field]

on cmd

it wont open😭

Idk maaan

[glacial pollen]

lmao, bruh

[proper shale]

delete stored_setting.json, restart okada

[scarlet wedge]

How do I know if the overtrain point has already reached without having to wait a long time to see if that was the lowest point?

[scarlet wedge]

Do i have to look the g/mel and g/kl for this?

[proper shale]

lowest point ≠ overtraining after lowest point

it is very likely that the most recent model (unless in overtraining) would be the best

overtraining takes a long time...

[scarlet wedge]

ok

and what the g/mel and g/kl show?

[proper shale]

mel = how well it's reproducing clarity
kl = how different the model is from the dataset

lower is better

[scarlet wedge]

thanks

[proper shale]

yw :)

[granite falcon]

Anyone?

[ancient field]

how to del that help

[glacial pollen]

There's lots of ways

one is understanding how graphs work

and what's the relationship between one metric and another

Take these. Should help you ~ yw. (( ps. syncing is only available on my fork, mainline and other forks which do not corporate ' auto syncing ' nor are based on mangio releases

Other than that, will post a real life example of overtraining in a bit ( most likely

( during training rn

[jovial tulip]

Thank you, the previous problem was resolved, but now this one appears:

LINK : fatal error LNK1158: não foi possível executar 'rc.exe'
error: command 'C:\Program Files (x86)\Microsoft Visual Studio 14.0\VC\BIN\x86_amd64\link.exe' failed with exit code 1158
[end of output]

note: This error originates from a subprocess, and is likely not a problem with pip.
ERROR: Failed building wheel for fairseq
Failed to build fairseq
ERROR: Could not build wheels for fairseq, which is required to install pyproject.toml-based projects

[glacial pollen]

Try any of these

[ancient field]

GUYS HOW TO DEL THIS

[glacial pollen]

But i reckon the path one should do the trick

@jovial tulip

[glacial pollen]

You should go and ask in #🔍│help-w-okada as this channel is for rvc

[ancient field]

pls

[glacial pollen]

and not for w-okada

[jovial tulip]

I don't have these files in this path
Can I call you privately to send prints? I can't send it here in this chat

[rare goblet]

Ayo? @jovial tulip level 1 !!!

[jovial tulip]

Oh, now I can

[glacial pollen]

I don't really have much time rn for such things
however

you should do the env path thing first

rather than files related approach

oh, unless

wait, you don't have the files? did you correctly install the buld tools? @jovial tulip

[jovial tulip]

This path only has these files

[glacial pollen]

Hmmm... these are x86

and x64

it's for win10 sdk ver: 10.0.18362.0

so make sure it matches in your visual build tools

But yea, those are the files you want

@jovial tulip
That's all I can say and suggest really, I gotta go back to my project

Hope you can resolve the issue

Also, in any case, check out this:
https://stackoverflow.com/questions/43858836/python-installing-clarifai-vs14-0-link-exe-failed-with-exit-status-1158

[jovial tulip]

Oooh! Thank you very much! Now I did it! Thanks!!! @glacial pollen

[glacial pollen]

: ) yw

[granite falcon]

Ayo

[rare goblet]

Ayo? @granite falcon level 3 !!!

[proper shale]

close down okada first, it's inside the "MMVCServerSIO" folder or whatever you renamed it to

[slate compass]

ive tried both vac and vb audio

[stone pewter]

how can i fix sound cracks

[vagrant edge]

what are the best settings for 2000 credits?

[obtuse rapids]

how do i fix this @proper shale

[ivory patio]

How to make ai cover and wwhat interface link

[violet heron]

What’s your GPU

[rotund prawn]

Ask it to Gemini

[rotund prawn]

Clearer vocal

[obtuse rapids]

ummmmm

[rotund prawn]

No

That is a AI

Like gbt

[mystic pawn]

hey guys

[obtuse rapids]

ohhhhhhhhh ok

[mystic pawn]

does rvc disconected still works?

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• RVC Mainline, by Hina Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• EasyGUI, by rejects Google Colab
• UVR5 NO UI for Google Colab, by Eddy Google Colab

More commands: -audio, -hf, -local, -kaggle, -overtraining, -realtime, -rvc, -help

[rotund prawn]

There should be a Gemini option just below the error code.

[rotund prawn]

Yes

It works well

[mystic pawn]

Really? that's good to know thx

[rotund prawn]

Np

[rotund prawn]

If there is a problem
you can type me

[mystic pawn]

Thx dude

one last question, what was pretrain used for? i'm one of those old heads who used the colab a long time ago

but i don't remembar what was pretrain for

[rotund prawn]

What language your dataset

And how long

[mystic pawn]

English, it's around 1 minute

[rotund prawn]

Hmm

Titan could be best

Yes

TITAN

[mystic pawn]

That sounds good

[rotund prawn]

[obtuse rapids]

i can't fine it

[rotund prawn]

Wow

[wet parrot]

My voice doesnt change

when i use Itadori voice model

[rotund prawn]

What do you use

For ai cover

[wet parrot]

MMVC

[wet parrot]

ai cover?

[rotund prawn]

MMVC?

[wet parrot]

yes

[rotund prawn]

That program is not my area of expertise
It probably doesn't work with rvc

[wet parrot]

MMVCServerSIO

it does work with rcv

[rotund prawn]

The channel called help rvc

🤷

[wet parrot]

rvc*

[rare goblet]

Ayo? @wet parrot level 1 !!!

[wet parrot]

though i think i fixed it

[rotund prawn]

Good

[wet parrot]

oh nah

didnt fixed it

[rotund prawn]

Ss

Send ss

[wet parrot]

cant

oh

i can

well

programm name is W-okada

[rotund prawn]

I use it for realtime gen

[wet parrot]

thats why i need it for

oh wait there is channel for w-okada help...

fuck

[rotund prawn]

Bye bye

[brittle wing]

-colabs

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• RVC Mainline, by Hina Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• EasyGUI, by rejects Google Colab
• UVR5 NO UI for Google Colab, by Eddy Google Colab

More commands: -audio, -hf, -local, -kaggle, -overtraining, -realtime, -rvc, -help

[brittle wing]

-hf

[azure marsh]

<:huggingface:1179800228946268270> Hugginface Spaces

• 🆕 Ilaria RVC Zero, by thestingerx Huggingface Spaces
• 🆕 RVC⚡ZERO, by r3gm Huggingface Spaces
• Applio, by IA Hispano Huggingface Spaces
• AICoverGen, by r3gm Huggingface Spaces
• Advanced RVC Inference, by r3gm Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces

More commands: -audio, -colabs, -kaggle, -local, -overtraining, -realtime, -rvc, -help

[brittle wing]

Appello

[rotund prawn]

It's applio

[rare goblet]

Ayo? @rotund prawn level 14 !!!

[calm crown]

Why doesn’t it share a link of the gardio?

[brittle wing]

Bro I know I remember an Italian calling it appello

It sounds even better than applio

[rotund prawn]

Nevermind

[brittle wing]

Appello

[rotund prawn]

Italy scores own goal

[rotund prawn]

I love Spain too
But
I hope Italy wins
It looks impossible rn😭

[brittle wing]

Win what?

[rotund prawn]

Euro 2024

Italy vs Spain

Spain won😭

[celest falcon]

https://i.imgur.com/5m8t8sV.png

why do i get this error

[delicate kettle]

Does anyone know how to download audio conversion from kits.ai without paying for a subscription 😂

Cause I find it ridiculous that you have to do all of that now when it used to be free

[violet heron]

Do not use kits

More like shits.ai

Use RVC or applio instead for ai covers

[delicate kettle]

But no other sites have a good removing background vocals

💀

They actually do it good

[violet heron]

UVR/MVSEP:

(If you know the right settings)

[delicate kettle]

I don’t know any of that

[violet heron]

that’s why I’m telling you

[delicate kettle]

How am I supposed to understand UVR/MVSEP

[violet heron]

how are you supposed to understand literally anything else?

Learn it

It’s upload audio, select model and aggressiveness, and convert

UVR is local and MVSEP is online tho you will need a account for MVSEP (free)

[robust fractal]

like this @violet heron

[violet heron]

What program were you using

[robust fractal]

voice.ai

[violet heron]

don’t use voice ai

[robust fractal]

aaaa

[violet heron]

It’s not good

Are you trying to make covers or use it in realtime

[robust fractal]

so what apps do u recomend

[violet heron]

I’ll give you a better completely free alternative

[robust fractal]

both of it

[violet heron]

What’s your GPU?

[robust fractal]

RTX 3050

[violet heron]

For covers you can use RVC or applio

-local

[azure marsh]

🖥️ Local stuff

• 🍏 Applio, by IA Hispano GitHub
• Mangio-RVC-Fork, by Mangio621 Huggingface
• RVC Studio, by SayanoAI Huggingface
• AICoverGen, by SociallyIneptWeeb GitHub
• Replay, by Replay Team Website
• Original RVC, by the RVC-Project team GitHub
• GPT-SoVITS, by RVC-Boss GitHub

Credits to Faze Masta and Antasma for compiling these links.

More commands: -audio, -colabs, -overtrain, -realtime, -rvc, -help

[violet heron]

For real-time use w-okadas

-realtime

[azure marsh]

This interaction has expired, use the command /guides realtime if you wish to see it again.

[robust fractal]

whats the diff between 2 of em?

realtime and vovers

[violet heron]

Realtime is for using the voice in real time

Like as a voice changer

For covers it’s a ai cover

[violet heron]

-spaces Or if you don’t want to download RVC you can use Ilarias HuggingFace space which only takes about half a minute to convert a full song

[azure marsh]

<:huggingface:1179800228946268270> Hugginface Spaces

• 🆕 Ilaria RVC Zero, by thestingerx Huggingface Spaces
• 🆕 RVC⚡ZERO, by r3gm Huggingface Spaces
• Applio, by IA Hispano Huggingface Spaces
• AICoverGen, by r3gm Huggingface Spaces
• Advanced RVC Inference, by r3gm Huggingface Spaces
• RVC v2 Huggingface version, by Clebersla Huggingface Spaces

More commands: -audio, -colabs, -kaggle, -local, -overtraining, -realtime, -rvc, -help

[robust fractal]

aaaaaa okayyy okayy

i need to setup w-okada 1st i think

[violet heron]

They are both different programs so it’s just preference

[robust fractal]

i wanna use to talk with friends

like for fun

w-okada is prefer right?

[violet heron]

You would use w-okada for that

[robust fractal]

okay than thanks tim for the help

-realtime

[azure marsh]

This interaction has expired, use the command /guides realtime if you wish to see it again.

This interaction has expired, use the command /guides realtime if you wish to see it again.

[rare goblet]

Ayo? @robust fractal level 2 !!!

[hot lagoon]

on stacher.io How do I get the subtitles?

[brittle wing]

-colabs

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano Google Colab
• RVC Disconnected, by Kit Lemonfoot Google Colab
• RVC Mainline, by Hina Google Colab
• AICoverGen-WebUI, modded by Hina Google Colab
• AICoverGen-NoWebUI [English], by Ardha, fixed by Eddy, Hina and Gdr Google Colab
• AICoverGen-NoWebUI [Spanish], credits to Eddy, Hina and Gdr for translating and fixing Google Colab
• Ilaria RVC, by thestingerx Google Colab
• EasyGUI, by rejects Google Colab
• UVR5 NO UI for Google Colab, by Eddy Google Colab

More commands: -audio, -hf, -local, -kaggle, -overtraining, -realtime, -rvc, -help

[proud elbow]

unless u get subbed anime/movies from somewhere, try use Whisper (timestamp included)

[astral jungle]

Can give me the Text To Speech link?

[violet heron]

Applio and Ilaira RVC have tts or if you don’t want RVC you can use GPT sovits

[astral jungle]

Has anyone here tried the Training Model Tacotron2 yet?

[astral jungle]

Thank

[dull gazelle]

how do i read that? i mean how can i know if i should use 32k, 40k, 48k?

[thin stump]

thats 32k. Rarely will it be 40k

you'll know when it reaches 20khz

[dull gazelle]

but how can i know if its 48k?

[rare goblet]

Ayo? @dull gazelle level 3 !!!

[thin stump]

extremely rare so you can ignore that

[dull gazelle]

and if the video lenght is 2 minutes batch size per gpu should be higher than 4?

[thin stump]

I can't answer questions about vram since I don't have a gpu myself sorry. For short datasets like that you can use 2-4 if you would like

@olive cliff yeah i have no clue on how it works. I'll wait until you get a response

because i typed in http://blaise-tk/TITAN/resolve/main/models/medium/32k/ and that couldn't be right if it loaded so fast

[dull gazelle]

what if i use 25? i have 64Ram

does it gets better or no?

[thin stump]

higher batch sizes like that doesn't improve the training. You want to stick with power of 2's so for short datasets you can use a batch size of 4, default is 8, and 12-16 for longer datasets (30 min to an hour long)

[slate sinew]

is there like certain ai voice apps for different pc specs

if so i wanna know which one is the best

applio or rvc

idk bruh\

[rare goblet]

Ayo? @slate sinew level 2 !!!

[solid mantle]

does MMVC have tts option?

[glacial pollen]

easy, see where colors end?

around 16khz

In that program, you do *2

so that's approx 32khz

[dull gazelle]

is there a way to know if i am overtraining my model too?

[glacial pollen]

#✨│ai-help message

all it takes is a lil bit of scrolling or search my dude

¯_(ツ)_/¯

[gritty musk]

how do i make a coverrr

[carmine hearth]

https://docs.aihub.wtf/essentials/how-to-make-ai-cover/

[slate sinew]

where the hell do i place the .index and .pth files

[rotund prawn]

what the hell u use

[pastel oak]

pth in assets -> weights

index in logs

[arctic flume]

hey i need help

please its urgent

please

[rare goblet]

Ayo? @arctic flume level 1 !!!

[arctic flume]

please

please

how to fix this

@acoustic thorn

master help me your fellow padwan needs help

please understand its urgent

@acoustic shale

[proper shale]

outdated colab, use the newer ones in https://docs.aihub.wtf/ - RVC > Cloud section

https://huggingface.co/spaces/TheStinger/Ilaria_RVC is also a good option to use as well, easy to use too

[proud elbow]

that outdated colab hasn't been taken down

[low shard]

I wish it did @acoustic thorn

[cold tinsel]

is 2070 super 8gb enough to train locally?

[violet heron]

Yeah

[proud elbow]

[cold tinsel]

thank you

[fathom reef]

Why is Ilaria giving me low quality voice clips when using inference?

[brave garnet]

RVC Colabs and Spaces

⠀

HuggingFace Spaces 🤗

⠀
Ilaria RVC
EasyGUI port with some improvements, by Ilaria.

RVC-HFv2
Applio port, by r3gm.

AICoverGen
AICoverGen port, by r3gm.

Advanced RVC Inference
Extended version of the GUI with advanced settings, r3gm.
⠀

Made with love by Ilaria 💖

[proven hill]

can you elaborate?

[fathom reef]

So basically, I put in the highest quality voice clip that I have as the input, processed it, and it came out super low quality.

These are my settings, by the way.

I pretty much used the default ones except for the pitch and index influence.

You know what, nevermind.

I'll figure it out on my own time.

[rich crag]

RVC v2 Disconnected

[tough fiber]

why 32k sample rate isnt appear when i choose v2

bruhhh

how can i merge models wihtout okada.

[proven hill]

@low shard

[low shard]

hello

[low shard]

Mmm, maybe could depend on the voice model, could you try with another?

[rotund prawn]

select v1
then select v2 again

[astral jungle]

Has anyone here ever trained the tacotron2 model?

[thorny juniper]

hi

[rare goblet]

Ayo? @thorny juniper level 1 !!!

[thorny juniper]

how can i delete model

i want delete it

[pastel oak]

you can upload a new one in that same slot