✨│ai-help

[proper shale]

the install links are the download links...

[swift thicket]

i know

[rare goblet]

Ayo? @swift thicket level 1 !!!

[brittle wing]

A rx 6900 XT is somewhat performs similiar to a rtx 3090 in gaming

[rare goblet]

Ayo? @brittle wing level 2 !!!

[swift thicket]

but there are links for gpu and things like nvidia

i want the voice changer app

[severe violet]

Hey i have a question! how do i make the converted audio sound clean? in rvc? i've heard ai covers and they sound super clean! what's the trick here?

[proper shale]

yeah those are the download links for the voice changer, you select which one is the one for your gpu

[proper shale]

get studio vocals, if possible

and get a good model

[swift thicket]

https://github.com/w-okada/voice-changer/blob/master/README_en.md

is this the link?

[proper shale]

that's basically all you need for a decent cover

[proper shale]

no, that's just the readme

[swift thicket]

this?

[proper shale]

just use the links in the guide, for the type of GPU you havs

[proper shale]

yes.

[swift thicket]

i have nvidia

[ionic oak]

Yes but I did not find and understand what they explained to put the voice message on the RVC and then put the parameters

[severe violet]

How do i merge the vocals and the insterments? putting them on top of each other on audacity dosn't sound good! i

[proper shale]

yeah you need to download the file, put that in a folder, and do the steps in the GUI accordingly

[proper shale]

i think you can mixdown the tracks on audacity

[proper shale]

just read the page on what RVC thing you're using, like applio for example

[severe violet]

ye i'm not good with that!

[proper shale]

it should help you

[ionic oak]

In the GUI accordingly ? and what file ?

[rare goblet]

Ayo? @ionic oak level 2 !!!

[proper shale]

the voice file, aka your dataset

[severe violet]

can you link me up a youtube video that can help me with that?

[azure marsh]

📚 Audio Guides

• Perfecting Audio Isolation on Low-End Rigs: A Practical Guide, by Litsa The Dancer and Faze Masta Google Docs
• Gathering and Isolating Audio, by SCRFilms Google Docs
• Creating Datasets for RVC using iZotope RX, by LUSBERT Rentry
• Vocal Mixing Tutorial, by Roomie YouTube

🛠️ Tools

Audio Separation/Isolation

• https://mvsep.com/
• https://multimedia.easeus.com/vocal-remover/
• MDX-NET Colab

More commands: -colab, -uvr, -karafan, -overtrain, -help

[proper shale]

https://www.youtube.com/watch?v=rSxLF3kkBw0

this

[ionic oak]

Okay, but what exactly should I put in the .pth file and the .index?

[proper shale]

You don't have that yet, you're gonna train a model to get those

[ionic oak]

That's to say?

[proper shale]

You will use the voice message to train a model to get the pth and index

And then you can use the model

[ionic oak]

So I don't use this?

[proper shale]

Yeah no, this is for realtime voice changing. You'll use this later on, probably

Btw, I recommend asking your friend to send more voice messages, or audio files with their voice

19 seconds might not sound pretty good

[ionic oak]

I see, but what do I use and how do I test?

[proper shale]

You can use any RVC program, I suppose you installed original RVC since you have the realtime GUI.

Just open go-web.bat and you should be good to start trying to train

https://docs.aihub.wtf/rvc/local/mainline/

[ionic oak]

and then I follow this tutorial, right?

[proper shale]

Yup

[ionic oak]

yes but what do I do for the .pth and .index file, what do I put?

[proper shale]

skip that part, that's for when you have the model

look for the Training section

[ionic oak]

But before that I need several voice messages of approximately how long?

[proper shale]

Ask them for like, 5 minutes, can just be one file

Or multiple

5-10 minutes is already decent for a model

[uncut cairn]

i dont get how to install webui

[severe violet]

where is the audio guide channel?

[odd shale]

RVC webui?

[brittle wing]

Would you prefer a GTX 1080 or RTX 2060 for rvc

[severe violet]

how do i get to this guide

[uncut cairn]

yes

[rare goblet]

Ayo? @uncut cairn level 1 !!!

[odd shale]

You can read the docs.

-rvc

[azure marsh]

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

Full AI Voice Model Training Guide (Local)

Link: YouTube
Credits: Christopher Villanueva

Model training with Mainline RVC

Link: Rentry
credits: Raven (ravencutie21)

AICoverGen Colab Guide

Link: Google Docs
Credits: Eddy (Spanish Helper)

Create a model with RVC disconnected (colab)

Link: Google Docs
Credits: Angetyde

[uncut cairn]

ok daddy

[tame mica]

?

[severe violet]

can someone help me where do i find the audio guide in the discord server?

[odd shale]

I got the guide but it's outdated.

[brittle wing]

Do you guys somewhat think going from a RX 570 to a GTX 1660 super is somewhat a big leap in real time AI voice performance

Like less stutters and somewhat convincable voice tones

[tame mica]

https://rentry.org/RVC-dataset-RX-updated if you mean dataset guides

[severe violet]

what's the best process method for uvr5 vocals only?

[proper shale]

mmm maybe not

[proper shale]

BSRoformer

Has all vocals but

[stiff goblet]

does loss/g/fm work buggy? It always goes up after a certain step count

[proper shale]

You can get the leads with UVR BVE

[proper shale]

yeah it's kinda buggy

dunno why

the graph at least

[brittle wing]

What about a rtx 3060 ti

[stiff goblet]

Yeah, the other graphs works fine

[brittle wing]

I have the money to upgrade

[proper shale]

Mmm that would be very decent

[proper shale]

for now focus on mel and kl

[proper shale]

https://github.com/TRvlvr/model_repo/releases/download/uvr_update_patches/UVR_Patch_4_14_24_18_7_BETA_full_Roformer.exe

install this if y haven't yet

UVR 5 with the BSRoformer models

[severe violet]

you mean these?

[proper shale]

BSRoformer should be in the Download Center, after u download it it should be in VR architecture iirc

[west fox]

how can i fix this?

[severe violet]

i can't find it

[proper shale]

ye cuz you have to download it first, as i said

[severe violet]

yes but i can't find it in the download list

[rare goblet]

Ayo? @severe violet level 3 !!!

[proper shale]

also, make sure you're using that UVR version i linked

[severe violet]

oh! okay that make sence i didn't download it cuz the one i have and the one u linked have the same size so i was like they might be the same

[severe violet]

the thing is that's a ptch file how do instal it tho?

[severe violet]

how do i instal it? should i remove the old one or overwrite it or what?

ohhhhhhhhhh

it's not under the vr arch

let me check

nopppe i don't have it! they all start with mdxnet i don't have that one

[proud elbow]

install it first, no need to remove existing one

and then get BS roformer 1296 and 1297

[severe violet]

from the download center right?

and these models are for vocals right?

thank you let me install it

[primal herald]

umm

how to fix this problem

[rare goblet]

Ayo? @primal herald level 1 !!!

[primal herald]

---

JSONDecodeError Traceback (most recent call last)
<ipython-input-13-87685dcb27d0> in <cell line: 31>()
31 if os.path.exists(config_path):
32 # File exists, proceed with creation of creds and client
---> 33 creds = Credentials.from_service_account_file(config_path, scopes=scope)
34 client = gspread.authorize(creds)
35 else:

5 frames
/usr/lib/python3.10/json/decoder.py in raw_decode(self, s, idx)
353 obj, end = self.scan_once(s, idx)
354 except StopIteration as err:
--> 355 raise JSONDecodeError("Expecting value", s, err.value) from None
356 return obj, end

JSONDecodeError: Expecting value: line 1 column 1 (char 0)

[proud elbow]

-colab

[azure marsh]

Suggestions for @primal herald

☁️ Google Colabs

• Applio, by IA Hispano 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab
• 🆕 UVR5 NO UI for Google Colab, by Eddy Google Colab

How to use?

🤗 Hugginface Spaces

• Applio, by IA Hispano Huggingface Spaces
• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HFv2, by r3gm 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, -local, -overtrain, -realtime, -rvc, -help

[proud elbow]

use the above link instead of your outdated colab

[limpid plinth]

guys, i'm training on Applio online via browser. it does not create the index file in the collab folder. where it should usually be? and also, what could be the error?

the PTH file loads normally

[rare goblet]

Ayo? @limpid plinth level 2 !!!

[severe violet]

i have a question! is there any settings should i adjust in thervc web pageN to make the converted audio sound better

[primal herald]

thx helped a lot

[limpid plinth]

this one?

yes

via browser

[severe violet]

i have a question! is there any settings should i adjust in thervc web pageN to make the converted audio sound better

like a specific asetting that improves the quality

[rare goblet]

Ayo? @brittle wing level 51 !!!

[severe violet]

and can someone show me how do i mix vocal and instremnts so they sound like a one song!

how do you mix your vocals with insterments?

so you just put the vocal over the insterments no settings to ajdust? cuz i've heard ai covers in youtube and they sound better then mine that's for sure!

[limpid plinth]

it doesn't upload automatically, i need to upload it first. but as you can see, there's no file to upload

[severe violet]

i like to do things on my own! but thanks anyway! i'll find a video that can help me mix these stuff xD thanks again!

[rare goblet]

Ayo? @severe violet level 4 !!!

[uncut cairn]

for vocal separation in the webui, i entered the path of the audio folder to be processed

but it always says theres no directory or wtv

[severe violet]

what's next?

tes

as you can see in the picture

should i use esamble mode and choose both of them?

do they work for insterments too?

or vocals only?

😮 i like the results that i'm getting

what's colab?

OKay i just need 2 things noww and i'm set

a good model for insterments and how to mix and i'm good to go

background music

[earnest ivy]

what collab does everyone uses that works? need one for making ai cover

[azure marsh]

Suggestions for @earnest ivy

☁️ Google Colabs

• Applio, by IA Hispano 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab
• 🆕 UVR5 NO UI for Google Colab, by Eddy Google Colab

How to use?

🤗 Hugginface Spaces

• Applio, by IA Hispano Huggingface Spaces
• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HFv2, by r3gm 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, -local, -overtrain, -realtime, -rvc, -help

[earnest ivy]

thanks i'll try them out

[rare goblet]

Ayo? @earnest ivy level 1 !!!

[limpid plinth]

this is the local one, i tried with browser it said smt similar

no

ikik

before i could train only the index file but not the pth now it's the opposite

[severe violet]

stuck in audio mixing can someone give me a guide on how to?

[solid garden]

somone can help me in this errror on mangioRVC 23.7.0 local

recent call last):
File "E:\Mangio-RVC-v23.7.0\my_utils.py", line 96, in load_audio
ffmpeg.input(file, threads=0)
File "E:\Mangio-RVC-v23.7.0\runtime\lib\site-packages\ffmpeg_run.py", line 325, in run
raise Error('ffmpeg', out, err)
ffmpeg._run.Error: ffmpeg error (see stderr output for detail)

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "E:\Mangio-RVC-v23.7.0\infer-web.py", line 292, in vc_single
audio = load_audio(input_audio_path0, 16000, DoFormant, Quefrency, Timbre)
File "E:\Mangio-RVC-v23.7.0\my_utils.py", line 103, in load_audio
raise RuntimeError(f"Failed to load audio: {e}")
RuntimeError: Failed to load audio: ffmpeg error (see stderr output for detail)

File "E:\Mangio-RVC-v23.7.0\infer-web.py", line 316, in vc_single
audio_opt = vc.pipeline(
File "E:\Mangio-RVC-v23.7.0\vc_infer_pipeline.py", line 564, in pipeline
self.vc(
File "E:\Mangio-RVC-v23.7.0\vc_infer_pipeline.py", line 450, in vc
(net_g.infer(feats, p_len, pitch, pitchf, sid)[0][0, 0])
RuntimeError: CUDA error: the launch timed out and was terminated
CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect.
For debugging consider passing CUDA_LAUNCH_BLOCKING=1.
Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.

And this one

[sterile hinge]

would you guys say that the rvc disconnected or the local are the best way to train a model?

[rare goblet]

Ayo? @sterile hinge level 1 !!!

[proper shale]

you probably linked the path to a file and not a folder?

[proper shale]

and this, im not sure

[proper shale]

if you have a good gpu, go for local

[fierce isle]

is this outdated yet ive been using it for a good couple months. Is it worth my time to learn about the most updated RVC?

[severe violet]

Can someone help me with audio mixing please?

[proper shale]

Mmmm probably, you'd get faster conversions and more options

Since you're looking for something easy, I suggest using Applio, which is easy to use and has all the options you need.

https://docs.aihub.wtf/rvc/local/applio

[sterile hinge]

what we consider a "good cpu"?

[proper shale]

good gpu? at least like a 2060/2070

[sterile hinge]

exactly what I have. Afraid that will make the training not be that good

[proper shale]

That's pretty good for training tbh. It ain't the best experience but certainly better than training on 4GB VRAM

[sterile hinge]

Thanks

[proper shale]

You're welcome :)

[fierce isle]

ty

[proper shale]

You're welcome :)

[proper shale]

What's up?

[severe violet]

I just want a guide on how do I mix the concerted audio with the instrumental cuz if I just put them together they sound awful

[proper shale]

I believe I sent you one earlier

-audio

[azure marsh]

📚 Audio Guides

• Perfecting Audio Isolation on Low-End Rigs: A Practical Guide, by Litsa The Dancer and Faze Masta Google Docs
• Gathering and Isolating Audio, by SCRFilms Google Docs
• Creating Datasets for RVC using iZotope RX, by LUSBERT Rentry
• Vocal Mixing Tutorial, by Roomie YouTube

🛠️ Tools

Audio Separation/Isolation

• https://mvsep.com/
• https://multimedia.easeus.com/vocal-remover/
• MDX-NET Colab

More commands: -colab, -uvr, -karafan, -overtrain, -help

[proper shale]

https://www.youtube.com/watch?v=rSxLF3kkBw0

^^ this

[severe violet]

Ye I saw that video he is using paid plugin cal vocal

[severe violet]

Anyone else?

[proper shale]

Don't think you need that, though

You can always find free plugins for vocal stuff

I don't really do mixdowns professionally, since I only use a phone lol 💀

[proper shale]

Hmmm, you can find other vocal mixdown guides, lots of good stuff on YT

I don't have any specific recommendation, but... yeah, some of em should apply to AI stuff

[severe violet]

I think I'll just learn audio editing and mixing from scratch

[tropic trail]

-rvc

[azure marsh]

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

Full AI Voice Model Training Guide (Local)

Link: YouTube
Credits: Christopher Villanueva

Model training with Mainline RVC

Link: Rentry
credits: Raven (ravencutie21)

AICoverGen Colab Guide

Link: Google Docs
Credits: Eddy (Spanish Helper)

Create a model with RVC disconnected (colab)

Link: Google Docs
Credits: Angetyde

[tropic trail]

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab
• 🆕 UVR5 NO UI for Google Colab, by Eddy Google Colab

How to use?

🤗 Hugginface Spaces

• Applio, by IA Hispano Huggingface Spaces
• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HFv2, by r3gm 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, -local, -overtrain, -realtime, -rvc, -help

[solid garden]

yes, solved this

[solid garden]

but not this

[viral osprey]

How do I start

[proper shale]

ye im not sure why that's happening

[proper shale]

start what exactly?

[viral osprey]

Using Ai voice

Like idk where to start

[brittle wing]

Which one sounds better - a single pitch extraction algorithm or a hybrid one?

?!?

fcpe+RVMPE?

Without it being mixed, why?
Isn't an ensemble/mix/hybrid more convenient or makes it worse?
(Probably does)

RVMPE combined with other ones?

Have you ever used Applio?

It has a hybrid method

Hm are you sure?

Cause I probably noticed the same thing hybrids sound worse.

You're even a model maker so you know what you're talking about.

I do too

It has the best vocal range

[maiden remnant]

how do you make the voice models?

[turbid owl]

it is indeed fine without the index, i finally got the model to run using an actual rvc instance

[turbid owl]

pretty sure thats what #📑│making-models is for

[maiden remnant]

i meant train idk why i said make

[maiden remnant]

can i even train the ones i download from vocie models

too bad

[violet heron]

Those are already trained

[sweet ivy]

The only thing you can possible do is Merge it with another? but idk if that is good or bad.

[uncut cairn]

why is the vocal separation not working even though i entered the audio folder path 9000 times?

im done i did it 20 times

[violet heron]

1. You need the UVR models

And 2. Use UVR standalone or MVSEP for audio separation

[maiden remnant]

UVR is good

[rare goblet]

Ayo? @maiden remnant level 4 !!!

[uncut cairn]

are there any documents that explains all of these?

[rare goblet]

Ayo? @uncut cairn level 2 !!!

[glass stratus]

-rvc

[azure marsh]

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

Full AI Voice Model Training Guide (Local)

Link: YouTube
Credits: Christopher Villanueva

Model training with Mainline RVC

Link: Rentry
credits: Raven (ravencutie21)

AICoverGen Colab Guide

Link: Google Docs
Credits: Eddy (Spanish Helper)

Create a model with RVC disconnected (colab)

Link: Google Docs
Credits: Angetyde

[rare goblet]

Ayo? @glass stratus level 1 !!!

[glass stratus]

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab
• 🆕 UVR5 NO UI for Google Colab, by Eddy Google Colab

How to use?

🤗 Hugginface Spaces

• Applio, by IA Hispano Huggingface Spaces
• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HFv2, by r3gm 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, -local, -overtrain, -realtime, -rvc, -help

[glass stratus]

-train

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab
• 🆕 UVR5 NO UI for Google Colab, by Eddy Google Colab

How to use?

🤗 Hugginface Spaces

• Applio, by IA Hispano Huggingface Spaces
• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HFv2, by r3gm 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, -local, -overtrain, -realtime, -rvc, -help

[low shard]

check our docs to se the bedt models to use https://docs.aihub.wtf

[sweet ivy]

Is there way to focus on a single Vocal and have that extracted? Especially when there are multiple in a track?

[proper shale]

Just use karaoke separation models to get the lead and background vocals. UVR BVE and MDX Karaoke get the job done

[mellow sigil]

Hi i was wondering if there is a goku latam voice model or if anyone has one

[proper shale]

There probably is, just search for it on #1175430844685484042 or weights.gg

[brittle wing]

Do gtx series card usually perform better than rx series card in both okado and rvc?

[proper shale]

GTX I'm not sure, but RTX maybe

[brittle wing]

A lot of people in this channel have said that rvc performs pretty fine on a gtx 1650

Like the most budvet card

Budget*

[mellow sigil]

Im on a galaxy s23 ultra lol maybe that's my problem but i can't seem to get to work the gui

I keep getting this error

---

ModuleNotFoundError Traceback (most recent call last)
<ipython-input-15-4de11472052f> in <cell line: 4>()
2 #@markdown Link the URL path to the model (Mega, Drive, etc.) and start the code
3
----> 4 from mega import Mega
5 import os
6 import shutil

ModuleNotFoundError: No module named 'mega'

I've tried about 3 models so far

[rare goblet]

Ayo? @mellow sigil level 1 !!!

[mellow sigil]

Any tips?

[olive hamlet]

Are there any ways to increase the performance of RVC? At the beginning of the game, for example, to emulate a streamer, etc., the value counter in res begins to exceed the buf values and because of this, the models hang heavily. I even tried to squeeze out more value for the system (so that it spends more resources)
my accessories. RTX 4070 TI and Intel(R) Core(TM) i9-12900F 12th generation processor

[queen kraken]

try pip install mega.py before starting the code. if it doesn't work, modify the code to from mega.mega import Mega

[compact sinew]

vc settings?

[olive hamlet]

I tried, but it didn't work out

[compact sinew]

show me ur voice changer settings

[cinder ivy]

(this appears when downloading the voice model in google colab)
JSONDecodeError Traceback (most recent call last)
<ipython-input-5-be56f2050b6f> in <cell line: 31>()
31 if os.path.exists(config_path):
32 # File exists, proceed with creation of creds and client
---> 33 creds = Credentials.from_service_account_file(config_path, scopes=scope)
34 client = gspread.authorize(creds)
35 else:

5 frames
/usr/lib/python3.10/json/decoder.py in raw_decode(self, s, idx)
353 obj, end = self.scan_once(s, idx)
354 except StopIteration as err:
--> 355 raise JSONDecodeError("Expecting value", s, err.value) from None
356 return obj, end

JSONDecodeError: Expecting value: line 1 column 1 (char 0)

[low shard]

outdated colab

use newer ones in our docs https://docs.aihub.wtf

[low shard]

Nope, your using colab, it uses a remote cloud pc gpu of google

[low shard]

are you using ilaria rvc colab? Js click runtime, disconnect and delere runtime, amd re run

[dense marsh]

-colav

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab
• 🆕 UVR5 NO UI for Google Colab, by Eddy Google Colab

How to use?

🤗 Hugginface Spaces

• Applio, by IA Hispano Huggingface Spaces
• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HFv2, by r3gm 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, -local, -overtrain, -realtime, -rvc, -help

[molten dirge]

Hey, I was curious if someone could help me with streaming. I use hina colab

[proper sapphire]

Anybody know the cause of this?

for some reason it looped back to training epoch 1 after epoch 840

[gusty hound]

you know who else looped back to training epoch 1 after epoch 840? MY MOM!

[proper sapphire]

oh

[proper shale]

I suppose something went wrong with the D/G files' saving process...

[proper sapphire]

it saved just fine

i'll just delete the most recent save

what

[proper shale]

😭 whattt

[proper sapphire]

yeahhh

so i had it save every 5 epochs

and i did 840 epochs

so basically

there were 168 files 4-8 gbs big

i needed to erase most of them

[pastel oak]

Yes you should be deleting them as you go with a low save epoch like that

[proper shale]

you should've enabled "Save only latest .ckpt" ig

[proper sapphire]

nah

[maiden remnant]

what female ai model is the most trained and free?

[violet heron]

More trained != better

[maiden remnant]

i meant whitch one is free that has been trained a lot

[violet heron]

More trained does not mean better

[maiden remnant]

wym

thats not logical

[violet heron]

There is such thing as overtraining

And best epoch count depends on the dataset that was used to make the model

[severe violet]

Any good ai automix tools?

[odd shale]

Nope.

Use programs like Audacity or FL Studio to mix yourself.

[odd shale]

Actually overtraining exists.

the ideal epoch count will depend on your dataset length and the voice you wanna train.

You must check Tensorboard to know when your model overtrains.

[maiden remnant]

whats ur pc specs?

[odd shale]

I have a 16GB Ram laptop.

But why that question?

[maiden remnant]

i want to build a 1000euro pc for ai

so i want to know whats good

i bough mine for the same price in 2018

but its not for ai

[fierce crown]

Where can i download ROD

uhm is spek supposed to look like this

[wheat hound]

is mmvc compat w/ rx7000+ ?

bec i see no difference with cpu/gpu0 options

max cpu power rises res upto 1400ms w/ a 1000 epoch 40k model

[proper sapphire]

is there any way to go beyond 1000 epochs

Nvm I've gone even further beyond

https://tenor.com/view/vegeta-super-saiyan-dbz-比達-達爾-gif-7248615

damn

[cinder ivy]

is there a way to use the local version without opening the .bat file cuz i have macOS

nvm

[rare goblet]

Ayo? @cinder ivy level 1 !!!

[teal tendon]

i have everything set up, my issue is it wont work in game, ive been trying to mess around with my input and outputs with no luck, i posted this in #🔍│help-w-okada by accident

[rare goblet]

Ayo? @teal tendon level 1 !!!

[hidden dew]

does anyone know how to use RVC on oracle cloud? if it even exists

[proper shale]

Probably doesn't

But maybe if you could load .ipynb files in there, it could work? You'd have to adapt some stuff but maybe it would work... not sure how Oracle Cloud works

[severe violet]

hello i have a question! what's the best model in uvr5 to extract background noises only? like echo and clapping and stuff like that

[proper shale]

Echo can be dealt with UVR De-Echo, other sounds could be dealt with either BSRoformer and Bandit Plus

https://docs.google.com/document/d/17fjNvJzj8ZGSer7c7OFe_CNfUKbAxEh_OBv94ZdRG5c/edit#heading=h.hyzts95m298o ---> best denoising models n stuff

https://colab.research.google.com/drive/1efoJFKeRNOulk6F4rKXkjg63RBUm0AnJ?usp=sharing ---> colab for Bandit Plus n stuff

[severe violet]

thanks for the reply! you didn't get what i'm trying to say! i don't want to remove those effects! i want insted of extracting vocals only or instermntal only! i want to extract those background effects only! i hope you get what i'm trying to say! thank you

[proper shale]

BSRoformer and Bandit Plus could probably get those though, so

Yeah

I dunno if it'll sound great though

[severe violet]

i have bsroformer installed i think these are for vocals and insermntal only? am i correct?

[proper shale]

Yes, but they can also deal with background noise...

You can also use Bandit Plus and UVR's De-Echo/De-Reverb stuff to get SFX and echo/reverb respectively

[severe violet]

thanks for the reply once more! all i want is to put background noise in audio file sperate! i want to use those noises! what do i do?

[rare goblet]

Ayo? @severe violet level 5 !!!

[proper shale]

Just use the Instrumental output (or the Vocals output if it's De-Echo/De-Reverb)

[severe violet]

Alright i'll just try every model xD! thanks for the help tho

[proper shale]

You're welcome! Good luck!!!

:)

[keen swift]

hi! i'm trying to download an avril lavigne model but i'm getting this

"JSONDecodeError Traceback (most recent call last)

<ipython-input-5-87c5b8b3162e> in <cell line: 31>()
31 if os.path.exists(config_path):
32 # File exists, proceed with creation of creds and client
---> 33 creds = Credentials.from_service_account_file(config_path, scopes=scope)
34 client = gspread.authorize(creds)
35 else:

5 frames

/usr/lib/python3.10/json/decoder.py in raw_decode(self, s, idx)
353 obj, end = self.scan_once(s, idx)
354 except StopIteration as err:
--> 355 raise JSONDecodeError("Expecting value", s, err.value) from None
356 return obj, end

JSONDecodeError: Expecting value: line 1 column 1 (char 0)"

i'm not sure what it means exactly, i appreicate if anyone could help!

[proper shale]

Seems like you're using an outdated Colab, use the ones in https://docs.aihub.wtf/ instead - guides are also there too.

Just check the RVC > Cloud section and be happy

[keen swift]

got it! only problem is i don't see three dots to download

wait nvm i figured it out, i only had to right click LOL

[rare goblet]

Ayo? @keen swift level 1 !!!

[proper shale]

ah

Classic Firefox

[keen swift]

ohhhh now i c the problem lmfao

tysm fellas

[proper shale]

You're welcome :)

[sweet ivy]

@proper shale thats how my graph looks. sry for posting here can't post pics in the other chat.

[proper shale]

Mmm feels like collapses

[sweet ivy]

mel and kl are still going down.. didn't flatten out yet

[proper shale]

That kinda hinders the learning process, but the model can still improve

[proper shale]

Yeah dw about it then

[sweet ivy]

kl did go negative here and there... is that a sign of collapse?

[mystic axle]

Started fresh from the exe compiled version, placed in my documents folder with no funny characters and not running in administrator mode. This is after running "run-applio.bat". No clue how to fix this 😭

[proper shale]

Try running it without administrator mode, perhaps

[proper shale]

Mmm that's pretty decent

[proper shale]

I have never had kl go negative

So uhhh

[mystic axle]

i have been to no aveil, ooop

this just ran

nvm didnt fix anything 😭

[proper shale]

oh no

[mystic axle]

i sometimes get different errors too when running the "run-applio.bat" 😭

[proper shale]

frankly im not sure what's up...

[proper shale]

hmmmmm

interesting insight... but not sure what's even causing that

[mystic axle]

thank you for taking a look, ill just prob fresh install from zip

what

bruh

💀

im so confused but wtv i guess 😭

[proper shale]

😭

im also confused

[mystic axle]

that was fucking weird

[proper shale]

but hey, it works!

[sweet ivy]

what is your python version?

[mystic axle]

3.11.8

[rare goblet]

Ayo? @mystic axle level 3 !!!

[sweet ivy]

torch 2.1 >=3.8, <=3.11

I am guessing you need python at or below 3.11 for it to function normally but if it works now... don't change anything... Don't fix it if it ain't broke or broke anymore

[mystic axle]

fs, thank you!

[proper shale]

3.9/3.10 are the best for RVC related stuff.. but idk if it's needed since it's Applio prepackaged

iirc 3.10 is the best one, but, regardless

[sweet ivy]

by 2029 we will have no compatibility issues anymore... we will just ask our local j.A.r.v.I.s. to rebuild the package for the latest python version and add a few features while at it 💀

[proper shale]

Neuralink

[cinder ivy]

Are there any apps that can extract the background vocals from the front (main) vocals?

[rare goblet]

Ayo? @cinder ivy level 2 !!!

[proper shale]

Apart from UVR? Not much

[cinder ivy]

Oh I tried using it but I got a little question abt it

[keen swift]

hi sorry again lol, any ideas on how to fix high pitches... not sure if that's the appropriate question here

[sweet ivy]

I used Spleeter a while back not sure if it was able to seperate vocals... let me try that right now

[cinder ivy]

Any way to fix that

[proper shale]

Had to delete cuz of instrumentals

But, I feel like that's just RVC being RVC

[cinder ivy]

I see

[severe violet]

Hello i have a question about the converted audio in rvc! sometimes it has a weird noise at some points like electricity noise idk how to describe it! even tho the vocal that i'm using are supppper clean! idk if there's some settings i need to qdjust! please help!

[sweet ivy]

Something i even struggle with here and there but i just keep moving the sliders back and forth until the staticky melows out. What settings are you currently using?

[severe violet]

what sliders? all of them?

[past slate]

Help with this, Hours ago I accidentally deleted my weights.gg account google, and I want to recover it, could you help me?

Hey

[thin stump]

Has anyone told you not to play with fire? 😭 You can't recover it

[past slate]

😞

[past slate]

Maybe I have to wait a few days to recover the account, right? because it says there that you deleted it recently, I mean just 😟

[thin stump]

This question has been asked on the weights discord server before and they can't recover what's deleted

so just make a new account

[past slate]

that wighets.gg screw all his big ass mother for a stupid opportunity

But ok, with my discord account, ITS permanently

[thin stump]

Nah you touched that button and let your intrusive thoughts win. Ig you can leave them feedback so they can make it dummyproof. Join the weights discord

you can leave a post under feature-request

[past slate]

I already know how to recover it, look for a shortcut. first you go to the playstore and look for the weights application, then when you enter the application you log in your google account and instead of rejecting it it accepts it, after that you go to the weights page, and then you don't care, start session in Google, and so you can skip that sign that the account no longer exists, what an epic shortcut xd

[thin stump]

lol ok if that works for you

[past slate]

Yes

Haz el intento xd, im mexican

[maiden remnant]

does it work on linux

the voice changer?

[uncut cairn]

how do i use pretrained models? i don't understand

[uncut cairn]

i got it

[proper sapphire]

Is this overtraining?

My g/total went up

but my mel goes down

also my kl

[dense marsh]

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab
• 🆕 UVR5 NO UI for Google Colab, by Eddy Google Colab

How to use?

🤗 Hugginface Spaces

• Applio, by IA Hispano Huggingface Spaces
• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HFv2, by r3gm 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, -local, -overtrain, -realtime, -rvc, -help

[rare goblet]

Ayo? @dense marsh level 5 !!!

[grizzled moon]

the colabs are really stubborn rn

i cant use inference

the output just goes error

[proper sapphire]

Anybody have a good site to seperate vocals from music?

for some reason the site i use doesn't recognize flute as music

[violet heron]

MVSEP

[proper sapphire]

MVSEP wasn't that great for me

[violet heron]

make a account tho otherwise the queues will be long

[violet heron]

Depends on the model

Or you can use UVR

But that requires a download unless you use something like google colab for it

[proper sapphire]

what model do you use

[violet heron]

Don’t share as it might be copyrighted

I would delete

[violet heron]

Depends on the song for me

[proper sapphire]

generally

[violet heron]

Karaoke 6 or whatever it’s called because it can someone’s separate backing vocals

Tho I would do testing yourself to see what sounds best for your songs

[uncut cairn]

For tensorboard, it says this in the case of rvc disconnect

it doesn't open up by itself though

[tame mica]

in rvcdisc there should already be a tensorboard in the cell

[uncut cairn]

nevermind im sorry i confused

i'm unable to see any models

[proud elbow]

refresh or wait till it logs one or more epochs

[uncut cairn]

oh

i have to click train first

its over for me

[olive hamlet]

I'm not on the computer anytime soon. But it seems like everything there is either standard or slightly modified

[rare goblet]

Ayo? @olive hamlet level 1 !!!

[compact sinew]

idk then if u dont show me i cant help u

[ashen cloud]

what do i do\

[olive hamlet]

can I send you a screenshot easier because I don't want to write a lot. There will be basic settings and advanced settings

[proper shale]

ok seriously why is there a folder that is only a space as its name

💀

[edgy swallow]

is there any voice that doesnt sound robot like at all

[crude bolt]

see #🔍│help-w-okada

[velvet geyser]

Do yall have any tips on what you use for cleaning up audio? I’ve been using vocal separation then noise cleanup but i still just get… a bunch of sound effects floating around

[rare goblet]

Ayo? @velvet geyser level 1 !!!

[crude bolt]

the best ofcourse have a clean source , so you wont have to bother cleaning up in the first place. I'd recommend UVR.

[proud elbow]

BS roformer or bandit plus in mvsep as first attempt, though some non-tonal sound effects are somewhat hard to remove. then do declick, manual cleaning and denoising.

[brittle wing]

What’s the best way to train vocals with pre-trainers?

[proper shale]

custom pretrains, you mean?

you probably don't need them, unless ur model doesn't sound right with original pretrains

[proud elbow]

you can try some in #1235952130855010365 , or simply use default/original one if not sure

[crude bolt]

a pre-trainer is merely used so you save time.
With a pretainer it doesn't have to figure out voices in the first place.

You can have the same or maybe better results depending the quality of the soundsamples, time and hardware to throw at it vs using available pretrainers.

I'd recommend the titan pretrain, if your computer can handle it ^^.

As an example: I wouldnt recommend you to use an english pretrainer, to create a voice-model for another language

[proper shale]

it helps on the development, but can have negative effects in the model, hence why original is probably the best, as it doesn't really fail like the others

[proper shale]

this is rvc realtime

[proud elbow]

most of them are finetuned from another one, including the original pretrain. can have some pros and cons according to experiments having been made so far.

[serene prairie]

is rvc or w-okada better?

I dont mind personal opinions

[proper shale]

Okada is easier to use overall, but can have problems, especially on AMD

[serene prairie]

is that what you use?

[proper shale]

yeah, it's the only one that supports the Apple Silicon stuff

[serene prairie]

fair enough

does it work witht he models under RVC? might be a dumb question buh ion know abt this stuff

[proper shale]

Yeah, it works with RVC models

And a bunch of others, but no one really uses those, like who would use SVC in realtime 😭

[serene prairie]

lol got it

[rare goblet]

Ayo? @serene prairie level 1 !!!

[serene prairie]

thanks for helping gang

[proper shale]

yw :)

[fleet hedge]

Why does the mangio rvc instantly error when I press "Convert"

[rare goblet]

Ayo? @fleet hedge level 1 !!!

[fleet hedge]

[proper shale]

check command prompt

[fleet hedge]

lalala

[proper shale]

something wasn't selected properly...

instead of using auto-detect path, copy its actual path and paste it above the dropdown, might be better

[fleet hedge]

Aaaa thanks!

[proper shale]

Did it work?

[fleet hedge]

Yup

[proper shale]

Nice! Glad I could help :)

[severe violet]

hello the converted audio playback speed in rvc is too fast! how do i slow it down or gmake it normal

[crude bolt]

Use an audio program that can up or downsample , squeeze and stretch audio files.

I would use FL (fruity loops) for it, but that is not free.
I often hear free wav editors like Audacity but i havent got much experience with it

[delicate sage]

how to make character sing a song i have the vocals and character model i want

[wheat cloak]

@proper shale When I imported a new voice from the models channel it did have an index file

[proper shale]

That's... good?

What's the issue though

[wheat cloak]

*Sorry didn't

@proper shale

[proper shale]

Ah

Shouldn't be such an issue though... you can use models without index

[wheat cloak]

Not for me

[rare goblet]

Ayo? @wheat cloak level 1 !!!

[wheat cloak]

It does not work

[severe violet]

hello the converted audio playback speed in rvc is too fast! how do i slow it down or gmake it normal

[dense marsh]

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab
• 🆕 UVR5 NO UI for Google Colab, by Eddy Google Colab

How to use?

🤗 Hugginface Spaces

• Applio, by IA Hispano Huggingface Spaces
• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HFv2, by r3gm 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, -local, -overtrain, -realtime, -rvc, -help

[dense marsh]

which ai voice changer should I use

[onyx phoenix]

so vocals + Character Model now what lmao?

[odd shale]

Use or RVC-Realtime or W-Okada.

-rt

[azure marsh]

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

[odd shale]

What do you mean?

Do you wanna make an AI cover or a voice model?

[onyx phoenix]

cover

[odd shale]

Welp, you can use MVSEP to isolate vocals.

First use BS Roformer to isolate the acapella

Then the isolated acapella pass it thru UVR BVE-44100 model on the same site to get the lead vocals

Unless the song is only lead singing

And finally, pass the isolated lead vocals thru the DeEcho-DeReverb model.

These 2 last models can be found under the Ultimate Vocal Remover HQ tab in the MVSEP site.

[brittle wing]

so i got about 1 minute of voice lines, how do i go about creating the 2 files needed for uploading in the voice changer

[odd shale]

1 minute of voicelines?

From which character?

You need a model to use in W-Okada.

[brittle wing]

its my friend

[odd shale]

You got 1 minute of audio of your friend?

[brittle wing]

yes

[proper shale]

What are you using?

[odd shale]

Actually you should ask him to record more audio of his voice.

[proper shale]

Right click and see if you can do anything

[brittle wing]

how many minutes should be enough

[odd shale]

at least 10, 15 or 20

[wheat cloak]

Kendrick Lamar

[odd shale]

what

Did you mean, are you using a Kendrick model?

[wheat cloak]

Yes

[proper shale]

No, like - what are you using to convert vocals

[wheat cloak]

RVC 2.0

[brittle wing]

once i have at least 10 mins of audio should i use a database creator

[odd shale]

You must isolate vocals first.

[odd shale]

Database?

[wheat cloak]

Ik I did

But no index file shows up when I use it

[brittle wing]

i guess im watching a wrong yt video

what should i do next once i got 10 mins of audio

[odd shale]

Read our docs about voice model making.

-rvc

[azure marsh]

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

Full AI Voice Model Training Guide (Local)

Link: YouTube
Credits: Christopher Villanueva

Model training with Mainline RVC

Link: Rentry
credits: Raven (ravencutie21)

AICoverGen Colab Guide

Link: Google Docs
Credits: Eddy (Spanish Helper)

Create a model with RVC disconnected (colab)

Link: Google Docs
Credits: Angetyde

[proper shale]

okay, but where? what fork?

[wheat cloak]

Fork?

[proper shale]

yeah, variations of rvc

[wheat cloak]

2.0

https://github.com/Tiger14n/RVC-GUI/releases/

This one

[proper shale]

outdated

[wheat cloak]

What should I be using?

[proper shale]

https://docs.aihub.wtf/rvc/local/applio/

this

applio

easy to use, has a lot of nice features

[wheat cloak]

Ty

[undone dove]

how do i train an ai voice

[spark gazelle]

how do i download the voice changer?

i have a gforce rtx 3060

ok thx

[rare goblet]

Ayo? @spark gazelle level 1 !!!

[undone dove]

how do i train a model

[lime tulip]

how long is crepe feature extraction supposed to take its been 5hrs already

[velvet geyser]

Yall dont reccomend UVR at all? I guess Im more asking advise for the best models, I use VOC FT, then ReverbHQ but I still get music behind the audio (obvi detecting it as vocals) What model do you reccomend in UVR for noise cleanup?

I can send an example here of a clip im having trouble with if that would help with the advise 🙂

I think most of my issue with the model im training is 1, lack of good gpu, anf 2, not enough clean sound so

(I have a 10 series rip)

[proper shale]

Use BSRoformer instead, you will need to download the UVR beta version: https://github.com/TRvlvr/model_repo/releases/download/uvr_update_patches/UVR_Patch_4_14_24_18_7_BETA_full_Roformer.exe

Should be in the Download Center once you've downloaded n all

Use either viperx 1296 or 1297

[velvet geyser]

Okay! Just run it and itll pop up to download? Ill try it out now!

[proper shale]

https://docs.aihub.wtf has everything you need to know about RVC and how to train

[velvet geyser]

Oh hell yeah, I was using uhhh

[proper shale]

The same as UVR normally is, tho I don't really remember how that goes

[proper shale]

if you have a large dataset on low crepe hop, yeah youll have to wait an eternity

[velvet geyser]

https://docs.google.com/document/d/13ebnzmeEBc6uzYCMt-QVFQk-whVrK4zw8k7_Lw3Bv_A/edit#heading=h.bjzhhhcn3f69

This is the tutorial I was using, is this one outdated?

Its actually pretty small, Only 7 minutes 😦

[proper shale]

Oh man 😭

This gives me nostalgia

Yes it's kinda outdated

[velvet geyser]

Character just doesnt talk that much and its like... impossible to get any clean audio outside youtube clips

which as far as im aware sucks

Okay!! Thats good to know! Its a lot but I think Im sort of getting the hang of it- my 10 series begs to differ though

[proper shale]

Ah dw it's better than being with an AMD card

You'd have to install Linux, get ROCm working...

[velvet geyser]

Oh I hate AMD Id never buy a card from them

LOL

Theyre not terrible but its just not nvidia

Thanks so much for the tips though!! Ill mess around with all this stuff 🙂

[rare goblet]

Ayo? @velvet geyser level 2 !!!

[proper shale]

Yw :)

[lime tulip]

its only like 4 minutes at 64hop

[rare goblet]

Ayo? @lime tulip level 1 !!!

[proper shale]

might be crepe bugged out

[lime tulip]

it seems to be running not frozen/crashed

[velvet geyser]

Crepe over RMVPE?

Even with a lower end gpu?

[lime tulip]

is it not recommended i did one 6min30sec rmvp model yesterday it took 14 hrs figured crepe would be faster than that sorry im new to this gpu is 3070 so pretty mid

[velvet geyser]

Sorry so yes crepe is better? Just dont understand your sentence LOL

[alpine bane]

thread:8, f0ing, Hop-Length:64: 50%|███████████████████████ | 2/4 [2:48:11<2:48:11, 5045.85s/it]

What does this "2:48:11<2:48:11" mean?

[spare mirage]

I need help

[proud elbow]

you might have looked for outdated guide. voc FT causes 17.5 khz spectrogram cutoff and there are other newer and better ones including that I mentioned before.

[velvet geyser]

RVC- its a program, This guy has a great tutorial https://www.youtube.com/watch?v=hB7zFyP99CY&list=PLknlHTKYxuNscCoVK-PQceoQO-TlLes8a

Or that lol

[lime tulip]

would also like to know

[velvet geyser]

bro what are you talking about its a python script run directly from your computer

[rare goblet]

Ayo? @brittle wing level 1 !!!

[velvet geyser]

Oh then yes collab lol

[thorny dock]

I want an easy way to do the voice overs like kits.ai

[velvet geyser]

If you want an easy way its not going to sound good

[thorny dock]

Thiz rvc seem 2 complicated

[velvet geyser]

Put some work in boy lmao

[thorny dock]

Actually I made a couple decent songs with kits b4 they charged

Eazy az fuc 2 use

[velvet geyser]

"2 complicated" = "i dont want to learn how to do anything myself" u can just be honest LOL

[thorny dock]

I could train a voice in Minutes

[velvet geyser]

Then pay em 🙂

[thorny dock]

Ahhh no

[velvet geyser]

Its not

Lol okay bye then!

No one giving you handouts goofy this is the help chat

[thorny dock]

Ok then whats a better way that's easy

[azure marsh]

Suggestions for @thorny dock

☁️ Google Colabs

• Applio, by IA Hispano 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab
• 🆕 UVR5 NO UI for Google Colab, by Eddy Google Colab

How to use?

🤗 Hugginface Spaces

• Applio, by IA Hispano Huggingface Spaces
• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HFv2, by r3gm 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, -local, -overtrain, -realtime, -rvc, -help

[velvet geyser]

byeeeee

[thorny dock]

Im juz looking 4 a free way 2 do thiz

[proud elbow]

wdym hacked? or u mean copyright issues?

[velvet geyser]

As far as manual cleaning, are you talking like-

Going into audactiy and something similar or?

Sorry for all the questions, I think this is the part im struggling the most with, bc a lot of the voice clips I have have a vocal audio track behind them thats getting picked up by BS Roformer

[proud elbow]

u mean backing vocals, or people talking in background?

[velvet geyser]

And denoising doesnt detect the vocals as noise of course

Um so like... music vocals. Playing behind the people talking, so I guess people talking in the bg?

We're telling you what to do, youre just telling us no bc you think its going to hack your computer, I dont know what you want man 😭

[rare goblet]

Ayo? @velvet geyser level 3 !!!

[velvet geyser]

I have no idea what that sentence means bro what

Ohhhh god youre a minor, nvm

[rare goblet]

Ayo? @brittle wing level 2 !!!

[median root]

any1 got a good beginner tutorial

[rare goblet]

Ayo? @brittle wing level 3 !!!

[brittle wing]

https://tenor.com/view/se-acabo-fin-se-termino-ya-basta-basta-gif-17180704

[viscid moss]

CoverGen takes a while, lemme try rq que colab

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab
• 🆕 UVR5 NO UI for Google Colab, by Eddy Google Colab

How to use?

🤗 Hugginface Spaces

• Applio, by IA Hispano Huggingface Spaces
• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HFv2, by r3gm 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, -local, -overtrain, -realtime, -rvc, -help

[viscid moss]

https://tenor.com/view/anya-forger-anya-spy-x-family-anime-salute-salute-gif-okey-dokey-gif-25700022

[dense marsh]

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab
• 🆕 UVR5 NO UI for Google Colab, by Eddy Google Colab

How to use?

🤗 Hugginface Spaces

• Applio, by IA Hispano Huggingface Spaces
• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HFv2, by r3gm 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, -local, -overtrain, -realtime, -rvc, -help

[feral talon]

hi! for my final i have to make a music video, we wrote lyrics but i wanted to use ai for the singer, and we found a free beat on youtube. am i in the right spot?

[leaden locust]

I keep getting this errro pls help

ModuleNotFoundError: No module named 'mega'

[proud elbow]

outdated colab, try use a link below

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab
• 🆕 UVR5 NO UI for Google Colab, by Eddy Google Colab

How to use?

🤗 Hugginface Spaces

• Applio, by IA Hispano Huggingface Spaces
• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HFv2, by r3gm 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, -local, -overtrain, -realtime, -rvc, -help

[leaden locust]

thanks

[proud elbow]

try sing it or have someone else sing it then convert it with an RVC model, or else, make vocal notes using vocaloid or synthesizer V

[feral talon]

would a voice model be able to do what i want? if not do i just search up vocaloid or synthesizer on google?

[hazy citrus]

hey I was wondering if anyone can help it's not connecting to my discord even after downloading the audio cable

[violet heron]

Send a screenshot of your w-Okada settings

[hazy citrus]

dms?

[rare goblet]

Ayo? @hazy citrus level 1 !!!

[violet heron]

Why can’t you send here?

[hazy citrus]

thought I needed perms

[hazy citrus]

I have my audio cable installed aswell as everything I needed to in the video I have watched

still somehow can't use it in discord voice chat's

[velvet geyser]

Yes, but you need the audio of a person singing first, to replace their voice, if that makes sense?

[rare goblet]

Ayo? @dusk knoll level 1 !!!

[sharp zodiac]

how do i set up this version of rvc, i use nivida btw

i am very familiar with rvc but i misplaced the file during a cleanup, and deleted it.

it was an old version from early 2023

[thin stump]

@brittle wing When do you know when to stop using the dc offset? Like one pass is enough?

[proud elbow]

if you have singing skill issue, vocaloid, synth V, or an auto tune plugin (Antares might be good one) could help you

[obtuse stream]

anyone help me? whats the different? Embedder Model
hubert or contentvec

[harsh bay]

How can I make an song using an YouTube existing song instrumental but whit my own lyrics?

[brittle wing]

I keep these parodies of people taking already existing songs and changing some certain lyrics whilst it sounds in tune with the lyrics/instrumental, anyone know how this is done?

[rare goblet]

Ayo? @brittle wing level 2 !!!

[brittle wing]

HuBERT is more general-purpose, aimed at capturing rich linguistic information from speech.

Whereas ContentVec is specialized for tasks that require separating the content of speech from the speaker's identity.

Depends on your needs.

ContentVec is generally better for singing.

HuBERT is great for talking.

[maiden remnant]

what does it take to make the best female ai voice model

like how u make that shi

where u can even whisper

[velvet geyser]

im about to lose my mind

Using Roformer Viper 1297, all of a sudden its erroring out at 2% and throwing up this log

(Copy pasted into notepad so yall dont die from spam)

why tf is it talking about denoise im useing roformer ugh

[slim geyser]

well to make a model that can whisper you need a lot of whispering data, which tends to screw over the other data you have in the process. Like you can only really whisper if you want to do that with the model, nothing else.

[rare goblet]

Ayo? @slim geyser level 23 !!!

[obtuse stream]

thanks 🙂

whats this?

[maiden remnant]

so why wont it work?

[brittle wing]

What RVC are you using?

[brittle wing]

What RVC are you using?

[maiden remnant]

this one

v2

[obtuse stream]

applio

[brittle wing]

Try "Mainline" it works for me, Applio and the rest of them have errors for me too.

[obtuse stream]

what is mainline?

[brittle wing]

What's your GPU?

Nvidia or AMD?

[maiden remnant]

mine amd

[brittle wing]

This is for AMD/Intel GPU users: https://huggingface.co/lj1995/VoiceConversionWebUI/resolve/main/RVC1006AMD_Intel.7z

[obtuse stream]

nvidia

[brittle wing]

This is for NVIDIA GPU users: https://huggingface.co/lj1995/VoiceConversionWebUI/resolve/main/RVC1006Nvidia.7z

[maiden remnant]

so i can actualy use my amd gpu for once

[brittle wing]

Enjoy :3

[maiden remnant]

https://cdn.discordapp.com/emojis/1184272969346850846.gif?size=48&name=BlobCatDance&quality=lossless

cant send gifs 😭

[brittle wing]

No worries, I clicked on it

uwu

[maiden remnant]

;p ❤️

[brittle wing]

Once you extracted Mainline, go to "go-web" and it should give you the interface.

Let's hope, yall's don't get any errors.

[maiden remnant]

nah if this works ama kiss ur feet

[obtuse stream]

ERROR: Exception in ASGI application
Traceback (most recent call last):
File "C:\Applio-3.2.0\env\lib\site-packages\uvicorn\protocols\http\httptools_impl.py", line 411, in run_asgi
result = await app( # type: ignore[func-returns-value]
File "C:\Applio-3.2.0\env\lib\site-packages\uvicorn\middleware\proxy_headers.py", line 69, in call
return await self.app(scope, receive, send)
File "C:\Applio-3.2.0\env\lib\site-packages\fastapi\applications.py", line 1054, in call
await super().call(scope, receive, send)
File "C:\Applio-3.2.0\env\lib\site-packages\starlette\applications.py", line 123, in call
await self.middleware_stack(scope, receive, send)
File "C:\Applio-3.2.0\env\lib\site-packages\starlette\middleware\errors.py", line 186, in call
raise exc
File "C:\Applio-3.2.0\env\lib\site-packages\starlette\middleware\errors.py", line 164, in call
await self.app(scope, receive, _send)
File "C:\Applio-3.2.0\env\lib\site-packages\gradio\route_utils.py", line 707, in call
await self.app(scope, receive, send)
File "C:\Applio-3.2.0\env\lib\site-packages\starlette\middleware\exceptions.py", line 65, in call
await wrap_app_handling_exceptions(self.app, conn)(scope, receive, send)
File "C:\Applio-3.2.0\env\lib\site-packages\starlette_exception_handler.py", line 64, in wrapped_app
raise exc

[rare goblet]

Ayo? @obtuse stream level 7 !!!

[brittle wing]

You're using Applio, correct?

[maiden remnant]

what are these chinese letters

[brittle wing]

Where at?

[maiden remnant]

whitch one is gpu

[verbal linden]

Hi all, I am new to this. is there any way I can use these voice models in colab? is there a notebook for these models? I found only for training RVC models, but not for use

[maiden remnant]

ask chatgpt if u dont get rresponse here

[brittle wing]

There's a better one, you don't need Colab

If you're only using Ai Models.

https://www.tryreplay.io/

[verbal linden]

its free?

[verbal linden]

Is it just for songs? I just wanted to voice the lyrics through these models

[rare goblet]

Ayo? @verbal linden level 1 !!!

[brittle wing]

Yes, it's free and Yes they have a Microphone option for you to speak/sing.

[verbal linden]

Does this program require a good computer? I have a weak computer

[brittle wing]

I don't think so, try it out.

[verbal linden]

Thanks, I'll give it a try

[proud elbow]

what spec? is it for training, inference or realtime voice changer?

[verbal linden]

I want to voice text. maybe I'll change the voicemails to a different voice

[proud elbow]

Applio has RVC and TTS integrated
#📰│dev-updates message

[verbal linden]

Thanks, I'll try it now.

[harsh bay]

I want to make an song using an artist voice on an instrumental whit custom lyrics how can I do that?

[rare goblet]

Ayo? @harsh bay level 1 !!!

[maiden remnant]

@brittle wing what do i do if i dont ahve an index?

[brittle wing]

What do you mean?

[maiden remnant]

when i load a model and i ahve only pth

but it wont work if i dont load index

[maiden remnant]

it dosnt work if i dont put in index

whitch i dont have

[brittle wing]

Is this your model?

[rare goblet]

Ayo? @brittle wing level 8 !!!

[maiden remnant]

nah

[proud elbow]

vocaloid, then convert voice using RVC if u want custom voice model

[brittle wing]

Try a different model that comes with it.

[maiden remnant]

it works like that

but dosnt with only one

so its like that?

[brittle wing]

Who ever publish that model probably forgot to add the index.

[wispy lodge]

You may use any model without index, it's not required

[maiden remnant]

seems like not

[proud elbow]

you can use a placeholder index file from another model, then set index rate 0

[maiden remnant]

i tried

[wispy lodge]

Choose any index file, but set index rate to 0

[maiden remnant]

but then the voice aint working

idk let me mess around a little

[wispy lodge]

Seems to work fine with wokada

[verbal linden]

I can write a text, and I can include emotion? laughter, sadness and all that. so that the model doesn't just voice my text, but also has emotion. to give humanity to the voice

[dull escarp]

yo how do i make the voice sound better

mine sounds trash ngl

[wispy lodge]

What exactly is wrong? It may depend on the model and on your settings

[dull escarp]

i keep picking up outside noises and the voice sounds strange

my input is already at 0.1

[wispy lodge]

You don't need to change input gain. If you have outside noises, you need to use noise suppresion options (sup1 or sup2)

[dull escarp]

where

[wispy lodge]

In wokada, they're here

[dull escarp]

my voice sound retarded fr

[rare goblet]

Ayo? @dull escarp level 1 !!!

[verbal linden]

how much this program weighs when downloaded to a PC?

[rare goblet]

Ayo? @verbal linden level 2 !!!

[wispy lodge]

Around 4-5gb I guess

[dull escarp]

anything else i need to change on it

[wispy lodge]

You need to increase extra. This improves quality, but increases GPU/CPU load

Also which GPU do you have?

[dull escarp]

where do i check

[wispy lodge]

In Task Manager > Performance tab

[dull escarp]

k

4060

[wispy lodge]

Then you downloaded wrong version, you need to get onnxgpu-cuda
https://huggingface.co/wok000/vcclient000/resolve/main/MMVCServerSIO_win_onnxgpu-cuda_v.1.5.3.18a.zip

It will work better

[dull escarp]

is that why it sounds bad

[rare goblet]

Ayo? @brittle wing level 1 !!!

[wispy lodge]

You need to follow the guide in install voice changer
https://rentry.co/VoiceChangerGuide

[rare goblet]

Ayo? @wispy lodge level 17 !!!

[wispy lodge]

That's one of the reasons. You can also check the guide, it includes recommended settings for your GPU

Like chunk and extra

[dull escarp]

which guide

[wispy lodge]

The one that I just linked

[dull escarp]

the onnxgpucuda is really big

lol

its a really big folder

[rare goblet]

Ayo? @cunning linden level 1 !!!

[proper shale]

just check the RVC > Local section, teaches you how to set up everything

either Applio or Mainline

[terse merlin]

What is the site you cover from?

[low shard]

if you wanna make ai covers, check our docs https://docs.aihub.wtf

[urban wasp]

-colab

[azure marsh]

☁️ Google Colabs

• Applio, by IA Hispano 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab
• 🆕 UVR5 NO UI for Google Colab, by Eddy Google Colab

How to use?

🤗 Hugginface Spaces

• Applio, by IA Hispano Huggingface Spaces
• Ilaria RVC, by thestingerx Hugginface Spaces
• RVC-HFv2, by r3gm 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, -local, -overtrain, -realtime, -rvc, -help

[proper shale]

Now just download the outputs, normally the first one should be the cover

Three dots > Download

[odd shale]

Download the outputs.

Do you have an program like audacity?

[proud elbow]

download or move that whole folder to your Drive

[odd shale]

Whenever you press on the three dots to download, a circle aside the file will appear.

[proper shale]

oh yeah good option too

Do that

[proud elbow]

drag to the directory drive/MyDrive/

[odd shale]

Try clicking on download.

[proper shale]

You need a virtual audio cable, so you can set it as your output in Okada and your input on Discord.

https://software.muzychenko.net/freeware/vac470lite.zip <--- VAC lite, one of the best virtual audio cables since it has no real issues

You're welcome :)

[wheat cloak]

Can I use RVC models with Applio?

[proper shale]

Yes

[dull escarp]

it still sounds bad even with those and i changed the stuff

[wheat cloak]

Do the models sound better with Applio then with RVC or wrose?

[rare goblet]

Ayo? @wheat cloak level 2 !!!

[wheat cloak]

Or the same

[proper shale]

Same thing

Both use RVC anyway

[dull escarp]

how do i make it sound better the voice sounds ass

[wheat cloak]

Ok

Tysm

[proper shale]

mess with tune and index ig

[proper shale]

You're welcome :)

[dull escarp]

also they can hear me talk first and then the ai voice goes

how do i fix it

[rare goblet]

Ayo? @dull escarp level 2 !!!

[wheat cloak]

[proper shale]

make sure you've got the virtual cable all set up, then just enable sup2 and echo options, increase s thresh

[dull escarp]

what is virtual cable

im just tryna talk on roblox

[proper shale]

It's what you need to use if you want your AI voice to go through

[proper shale]

_ _

mentioned this earlier

[dull escarp]

k

[proper shale]

just replace the input as the cable on whatever you're gonna use, and set it as output in okada

[static cedar]

Quick question. How can we clone our own voice

[dull escarp]

what do i do with the file

[proper shale]

open setup64, do the installing steps, should be all set up after that

[proper shale]

Record audio of yourself speaking or singing, and then just use RVC to do the training part.

https://docs.aihub.wtf/ has the more in-depth explanations - strongly recommend reading it.

[wheat cloak]

BRO WHAT DOES IT REALLY TAKE THIS LONG????

[proper shale]

😭 whattt

internet really acting up on applios download damn

[wheat cloak]

How long is usually take?

*Does it

[stuck kernel]

mac

[proper shale]

Really depends on the internet speeds you have, but...

Huggingface downloads shouldn't take that long? Or GitHub?

[wheat cloak]

I used huggingface

[proper shale]

Ye Huggingface gives me good download speeds...

[wheat cloak]

I was downloading a model at the same time though

So that download just finished so that should help it I think

[proper shale]

🙏

[wheat cloak]

Fr

I got basketball in a few hours so fr 🙏

[static cedar]

🙃I'm lost

[bright whale]

anyone have the link for colab?

[proper shale]

https://colab.research.google.com/github/iahispano/applio/blob/master/assets/Applio.ipynb

guides in https://docs.aihub.wtf/rvc/cloud/applio-colab

[proper shale]

Just check the model making section, and also the RVC essentials too

If you have questions about the docs feel free to ask

[bright whale]

Thank you, I'll try to remember how I do it, it's been a long time since I did a voice or song, if I have any questions, I'll ask.

[rare goblet]

Ayo? @bright whale level 1 !!!

[proper shale]

Alrighty :)

I'll be here to help you just in case

[bright whale]

I think the last time I did something like that was before discord was deleted that time

[proper shale]

Oh the good old days

[bright whale]

I should have reposted the models I had made, but I ran out of time lol

[rocky pecan]

@proper shale

[proper shale]

Hmmm... it's doing the step way too fast there

I think that model link might be broken

[rocky pecan]

damn :/

[rocky pecan]

i just tried another one and the same thing happened 🤔

[proper shale]

What's the Colab saying?

Wait... just realized - you do have an NVIDIA GPU, right?

[rocky pecan]

im not too sure

[rare goblet]

Ayo? @rocky pecan level 2 !!!

[rocky pecan]

how do i check

[proper shale]

Task Manager > Performance tab

[rocky pecan]

when i open the link to my model, it says this, is it normal?

[proper shale]

mmm, you probably have it privated then

[rocky pecan]

its not actually my model its someone elses

[proper shale]

or the person privated it,

[rocky pecan]

oh

those are the only two of the person i want :(((

[proper shale]

ah damn

you're out of luck then...

[rocky pecan]

[proper shale]

and btw you could probably run this on your PC

[rocky pecan]

thanks, i appreciate your help! 🙂

[proper shale]

You're welcome :)

[rocky pecan]

any idea what Error
Connection errored out. Failed to fetch means?

[proper shale]

mmm, apparently it didnt get the model?

[rocky pecan]

i uploaded it from a file

i downloaded a .zip

[obtuse cypress]

can someone tell me some webs that can seperate vocal (and free)?

[proper shale]

MVSep - Best online & free separator. I suggest creating an account for it so you get less queues.

[obtuse cypress]

Thank you so much!

[rare goblet]

Ayo? @obtuse cypress level 1 !!!

[proper shale]

You're welcome! :)

[minor zodiac]

Does anyone know how to overtrain a model?

Mine doesn't sounds like I expected

[minor zodiac]

https://vocalremover.org/

[sharp zodiac]

where is that?

[proper shale]

https://docs.aihub.wtf

Keep in mind you need a decent GPU.

[cinder relic]

hello i'm very new to this i'm trying my first ai cover but i keep getting this error code when trying to download the ai model off of the colab:

JSONDecodeError Traceback (most recent call last)
<ipython-input-10-404c6ad64c77> in <cell line: 31>()
31 if os.path.exists(config_path):
32 # File exists, proceed with creation of creds and client
---> 33 creds = Credentials.from_service_account_file(config_path, scopes=scope)
34 client = gspread.authorize(creds)
35 else:

5 frames
/usr/lib/python3.10/json/decoder.py in raw_decode(self, s, idx)
353 obj, end = self.scan_once(s, idx)
354 except StopIteration as err:
--> 355 raise JSONDecodeError("Expecting value", s, err.value) from None
356 return obj, end

JSONDecodeError: Expecting value: line 1 column 1 (char 0)

[cinder crag]

@wispy lodge - hi, thanks i tried Wokada and it works great besides the delay.. how low can i go with 3060 without erroring?

[rare goblet]

Ayo? @cinder crag level 3 !!!

[wispy lodge]

Recommended settings are here, but I think you can go even down to 64 chunk
https://rentry.co/VoiceChangerGuide#gpu-chart-for-known-working-chunkextra

[low shard]

outdated colab, use newer ones in our docs https://docs.aihub.wtf

[cinder relic]

thought that could be the problem, tysm

[cinder crag]

so if i get 4090 the ms delay will be significally better?

[brittle wing]

how do I uhh.. yknow

[wispy lodge]

Yes. You can also get better delay with modified version from my fork which may offer better performance
https://github.com/deiteris/voice-changer

[wispy lodge]

This is one of the screenshots with 4090 running this version from here #🔊│ai-development message

[cinder crag]

i'm a noob besides installing simple apps or using simple colabs.. is it more complicated to install your version?

[wispy lodge]

It's mostly the same, the installation instructions are included in the readme
https://github.com/deiteris/voice-changer?tab=readme-ov-file#how-to-use

[cinder crag]

so i just download another zip and do the same stuff?

[wispy lodge]

Yep

[cinder crag]

thanks! what does the VAC lite does?

[wispy lodge]

That's a virtual audio cable. Used to forward your audio to other applications

You could probably see VB-Cable as well, but VAC is the recommended one

[violet garnet]

hello ı have some issues

[rare goblet]

Ayo? @violet garnet level 1 !!!

[violet garnet]

some reason ıcant hear myself

in discord

ı even download vb

[proper shale]

Are you sure it's even converting your voice?

[violet garnet]

uuuuhhhh

wha?

[proper shale]

Set monitor to your headphones, that'll tell ya

[violet garnet]

ok let me try

ok ı set monitor headphones

[cinder crag]

can you estimate how many ms i could get in your version with my 3060?

[violet garnet]

now gonna work?

[violet garnet]

is gonna work?

[proper shale]

is it actually converting now?

if not, check command prompt for any errors

[violet garnet]

uuuuuuuuuuuuuuh

ıam turkey 😄

so...some words ı cant understand

[proper shale]

use translator

[violet garnet]

lol

ı did monitor thing

still not workin

can you....

screen share??

or somthing else like

ı can understand if ı cn see

[proper shale]

i said for you to do that so you can see if it's converting your voice, at all

is it though?

[wispy lodge]

I think chunk 32-40 (85-106ms) should be doable. GTX 1650 is capable of running at 40-48 (106-128ms), but that's basically maxing it out just for voice changer
#🔊│ai-development message

[violet garnet]

letme translate

[rare goblet]

Ayo? @violet garnet level 2 !!!

[violet garnet]

what can ı even do _

?

you saying "Are you sure it's even converting your voice?"

[proper shale]

yeah

is it?

[violet garnet]

ıam not sure...because ı cant even hear myself

[proper shale]

alright, so it isn't

now, check command prompt

[violet garnet]

ıdk how to use

[proper shale]

aka cmd

just send screenshots of what it says

[violet garnet]

the voice changer oor somthing else

?

[proper shale]

k so, 2 problems

replace crepe with rvmpe as your f0

and select your gpu in the gpu dropdown

[violet garnet]

did my voice changer crash???

like

ok wait let me open again

ok so

ı change my crepe to rvmpe

but

ı cant change gpu cause...

ı dont have any grafic cards

[proper shale]

mmmm

you do need a decent GPU to run it, so...

[violet garnet]

ı jsut realise somthing

how the f* one of the helpers is a TURKİSH!