✨│ai-help

[wary mantle]

Anyone know why I can't hear my voice changer client and it isn't working? Please I've been trying for A WHILE.

[sand iris]

The index file being renamed ?

[half cove]

No

[sand iris]

try #🔍│help-w-okada

[half cove]

Like both in a zip file

[sand iris]

ah yeah good catch

[quiet ridge]

Ok

[wary mantle]

what do I do there?

[sand iris]

ask there, it's the channel dedicated to it

[wary mantle]

ok thanks

[sand iris]

More eyes from people who know

[untold dome]

i need a zip link of the thing, like under #1175430844685484042, able?

[sand iris]

Apparently yeah

[half cove]

Download both and zip em then reupload it

[untold dome]

too much effort sers.. oh well

[pure knoll]

hi i need help

[pure knoll]

yo

[half cove]

Yea?

[modest spruce]

How do I get my model master back?

[rare goblet]

Ayo? @modest spruce level 5 !!!

[modest spruce]

I had it on the old server

[half cove]

You gotta apply

[violet heron]

You have it

[half cove]

"master"

[violet heron]

Using /jointeam

[modest spruce]

Nvm, found out my models “didn’t meet quality control” whatever the fuck that means.

If that’s the case idk why I’d even get model master in the first place.

[violet heron]

It means your models aren’t good enough

[violet heron]

They upped the requirements

[modest spruce]

What are they now?

[violet heron]

Should say

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

https://docs.aihub.wtf/v/brazil

🇪🇸 Spain

https://docs.aihub.wtf/v/spain/

🇫🇷 France

https://docs.aihub.wtf/v/france

🇩🇪 Germany

https://docs.aihub.wtf/v/germany

🇯🇵 Japan

https://docs.aihub.wtf/v/japan

🇰🇷 Korea

https://docs.aihub.wtf/v/korea

🇵🇱 Poland

https://docs.aihub.wtf/v/poland

🇷🇺 Russia

https://docs.aihub.wtf/v/russia

🇮🇹 Italy

https://docs.aihub.wtf/v/italy

[hoary zodiac]

how can i add my own voice model i made?

[violet heron]

To?

[hoary zodiac]

so people can "/search" it

[violet heron]

#1159514067187277865

Do /submit

[hoary zodiac]

ty

[violet heron]

-upload

[azure marsh]

huggingface.co upload Guides

English guide

Link: https://rentry.org/fdg_guide_newer
Credits: FDG

Spanish Guide

Link: https://rentry.org/AI_Hub_guide_esp
Credits: Julia (ailen2091)

Hindi guide

Link: https://rentry.org/AI_hub_guide_hi
Credits: Enes

Turkish guide

Link: https://rentry.org/AI_hub_guide_tr
Credits: Enes

[versed silo]

can someone help me my friend is able to hear my voice and the ai voice

[oblique venture]

do i need to train more? or am i overtrainning

[violet heron]

Once it gets to its lowest point

[oblique venture]

idk, it seem pretty flat here

my dataset is only like about 5 min

500 epochs

ah wait about 2 mins only

[thorny sorrel]

can I download models from kits.ai?

[hearty folio]

Guys RVC uses 50-70% cpu and below 30% gpu
is it normal or high cpu usage?

[fading blaze]

my client is lagging

and cant select gpu. it only gives me cpu option.

[thin stump]

No, you can't. Use weights.gg instead and #1175430844685484042

[fading blaze]

oops i figured out why. im using cuda instead of ml

[rare goblet]

Ayo? @fading blaze level 1 !!!

[brittle wing]

What do you guys use to isolate the main vocals? I tried BVE but Vocals only just did the music and insturmental didnt do a single thing, it just replayed the song back to me. I want JUST the main vocal

[obtuse nacelle]

So, what's the process like when trying to learn a voice?

[brittle wing]

Are you talking to me or asking your own question?

[fading blaze]

whats gpu 1 gpu 2 gpu 3?

[brittle wing]

It happened a third time. I am owed an explanation.

[rare goblet]

Ayo? @zinc ore level 1 !!!

[dawn nebula]

@brittle wing -realtime

-realtime

[azure marsh]

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

[vestal isle]

i need help

i dont know what do download to use the voicemodel i installed

[dawn nebula]

-realtime

[azure marsh]

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

[violet heron]

For covers or realtime?

[vestal isle]

realtime i have it but is it supposed to just work on videos or is it supposed to work in discord and games?

cause its only working in recordings

[violet heron]

-realtime

[azure marsh]

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

[violet heron]

Read the first one

[violet heron]

Discord and games

[icy roost]

hey i need help here

[violet heron]

Ask then

[icy roost]

so its specific. can i talk to you in dms

[rare goblet]

Ayo? @icy roost level 1 !!!

[icy roost]

hello?

[violet heron]

K

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

https://docs.aihub.wtf/v/brazil

🇪🇸 Spain

https://docs.aihub.wtf/v/spain/

🇫🇷 France

https://docs.aihub.wtf/v/france

🇩🇪 Germany

https://docs.aihub.wtf/v/germany

🇯🇵 Japan

https://docs.aihub.wtf/v/japan

🇰🇷 Korea

https://docs.aihub.wtf/v/korea

🇵🇱 Poland

https://docs.aihub.wtf/v/poland

🇷🇺 Russia

https://docs.aihub.wtf/v/russia

🇮🇹 Italy

https://docs.aihub.wtf/v/italy

[mellow stratus]

I m looking to build a voice cloning model. Can you help me in building the model and deploying on AWS.
Also, can you add these features:-

1. Translation from English to Spanish in same voice and pitch

[meager acorn]

dose it cost money to use the google colab

i launched it and its telling me to add coins of some kind

[acoustic hornet]

why is the robot saying this interaction failed

[timid herald]

how do i do this on mac?

[round plover]

how to do /search

[brittle wing]

hi can someone send me the model maker rvc link

[azure marsh]

Suggestion for @brittle wing

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

❔ English FAQ

https://docs.aihub.wtf/faq

Credits: AI HUB FRANCE

✨ More guides

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

How To Make an AI Cover With Ilaria RVC

Link: Rentry
Credits: 👽 Julia (ailen2091)

[thin stump]

Use RVC disconnected (colab) if you don't have a GPU

https://rentry.org/RVC_making-models

https://rentry.co/TrainingVoiceModels

The dataset part or gathering audio clips is just as hard because you have to isolate the noise from the background in UVR or MVSEP. Especially if there's no clean studio recording neatly compiled for you

[brittle wing]

guys what do I do in this step? Im confused and Im kinda dumb sorry

the dataset part

[brittle wing]

yes yes I know how to isolate

[brittle wing]

Am I have to upload the voice file to the "My drive" part?

aah okay! I made my model in the exact same way, months ago, It was in my language but I remember thank you so much! The yt video I was learning it with deleted all the links I thought it wasnt working anymore 😄

I dont have a dataset (the acapellas) yet, when will I do them, Ill start

thank you all verrry much

finally I can make sia sing sledgehammer 😄

[rare goblet]

Ayo? @quartz sky level 3 !!!

[hot lagoon]

overtrainning? or need to train more?

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

https://docs.aihub.wtf/v/brazil

🇪🇸 Spain

https://docs.aihub.wtf/v/spain/

🇫🇷 France

https://docs.aihub.wtf/v/france

🇩🇪 Germany

https://docs.aihub.wtf/v/germany

🇯🇵 Japan

https://docs.aihub.wtf/v/japan

🇰🇷 Korea

https://docs.aihub.wtf/v/korea

🇵🇱 Poland

https://docs.aihub.wtf/v/poland

🇷🇺 Russia

https://docs.aihub.wtf/v/russia

🇮🇹 Italy

https://docs.aihub.wtf/v/italy

[odd flower]

I want to add an accent to a model trained on my voice, accents like American accent or Japanese accent, is that possible if yes then how and how can we fix the shouting part?

[proper shale]

1. I'd recommend training the model with samples in these languages. If that ain't possible, just get as many variations as you can and disable index when inferring, then use samples from the language.

2. Don't use shouts for the dataset, I guess

[odd flower]

so let's say I am using a model that has an American accent downloaded from here, but I need accent for my voice

[proper shale]

Then keep using the index to get the American accent

There isn't a way to get a Japanese accent on a model like that unless you use a Japanese sample to infer

[upbeat sparrow]

i got a pth+index model here. what's the best model for tts with emotion?

[proper shale]

you can't really do TTS on RVC specifically, but you can do TTS elsewhere and convert the voice with RVC

I'd recommend ElevenLabs, has good emotion

[upbeat sparrow]

dont u have to pay for that?

[proper shale]

u get 10k characters for free every month

so idk

ig that should be enough for people

[upbeat sparrow]

oh

i can use my custom model?

[proper shale]

no

you'll have to use one of their voices, download the audio, then use RVC to convert

[upbeat sparrow]

ooohh

[odd flower]

okay

[molten pecan]

@feral sequoia #1159290752195633273

[tidal solar]

how do I pass the model link from weights.gg to rvc?

[proper shale]

download the zip, upload to huggingface and copy the download link

-huggingface

[azure marsh]

huggingface.co upload Guides

English guide

Link: https://rentry.org/fdg_guide_newer
Credits: FDG

Spanish Guide

Link: https://rentry.org/AI_Hub_guide_esp
Credits: Julia (ailen2091)

Hindi guide

Link: https://rentry.org/AI_hub_guide_hi
Credits: Enes

Turkish guide

Link: https://rentry.org/AI_hub_guide_tr
Credits: Enes

[proper shale]

there's ur tutorial

[compact quest]

how do i fix this problem

config\model \ transformer_Lm
creating build\lib.win-amd64-cpython-39\fairseq\config\model\wav2vec
copying fairseq\config\model\wav2vec\vq_wav2vec_gumbel.yaml -> build\lib.win-amd64-cpython-39\fairseq\config\model
wav2vec
creating build\lib.win-amd64-cpython-39\fairseq\config\model\wav2vec2
copying fairseq\config\model\wav2vec2\wav2vec2_base.yaml -> build\lib.win-amd64-cpython-39\fairseq\config\model\wa
v2vec2
copying fairseq\config\model\wav2vec2\wav2vec2_large.yaml -> build\lib.win-amd64-cpython-39\fairseq\config\model\w
av2vec2
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://visualst
udio.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

[sand iris]

it gave you a url

https://visualstudio.microsoft.com/visual-cpp-build-tools/

[compact quest]

okay

[tidal solar]

I have this error ---------------------------------------------------------------------------
BadZipFile Traceback (most recent call last)
<ipython-input-14-77e9d9bbba48> in download_online_model(url, dir_name)
52 print('[~] Extracting zip...')
---> 53 extract_zip(extraction_folder, zip_name)
54 print(f'[+] {dir_name} Model successfully downloaded!')

4 frames
BadZipFile: File is not a zip file

During handling of the above exception, another exception occurred:

Exception Traceback (most recent call last)
<ipython-input-14-77e9d9bbba48> in download_online_model(url, dir_name)
55
56 except Exception as e:
---> 57 raise Exception(str(e))
58
59 url = "https://hugging,example" # @param {type:"string"}

Exception: File is not a zip file

I followed the guide

[heavy mountain]

so, im training a model with about an hour and 33 minutes of samples, how many epochs would i use for it

[azure marsh]

Suggestion for @echo meadow

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

❔ English FAQ

https://docs.aihub.wtf/faq

Credits: AI HUB FRANCE

✨ More guides

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

How To Make an AI Cover With Ilaria RVC

Link: Rentry
Credits: 👽 Julia (ailen2091)

[proper shale]

Go for 500 in every dataset length

Though I do believe 1 hour is excessive... unless ofc there's a lot of variations

[heavy mountain]

the way i see it usually when i train, is that more samples = higher quality model at the end, which is why i used so many

[heavy mountain]

btw, how much is a dataset length, im not very experienced in model training, i just get samples and throw them in there, i dont really know what im doing too well

[proper shale]

yeah that's the dataset, im referring to the length of it because 1hr is excessive

[proper shale]

see, 30 mins would prolly be enough

if u have a good dataset it shouldnt be a problem to trim down

[heavy mountain]

alrighty, so 30 minutes and 500 epochs?

[proper shale]

yeah sure, go for that

[compact quest]

bro

i have AMD

i can't

[proper shale]

yeah uh, you can't train models on AMD cards

RVC stuff as a whole doesn't really work well in AMD cards

[compact quest]

fuck

[tropic cave]

I cannot use the audio file I want in rvc guide.

[delicate wharf]

So according to the training guide, should I be looking at Value or Smoothed?

[mental pivot]

What was the issue?

[proper shale]

Better to look at the actual value, but doesn't really matter a lot

[delicate wharf]

so just at the point the graph hit lowest and start raising again = overtraining, right?

[proper shale]

ehhhh, not necessarily

it's much more complicated than that

-overtrain

[azure marsh]

How do I know if my model is overtaining?

All-In-One Guide on how to make a good model

This guide explains how the D and G files works and much more: https://rentry.org/RVC_making-models

Credits: LUSBERT

Automated Overtraining Detection (AOD)

Will be available soon in #1159513888199540817

Credits: grvyscale

[proper shale]

this guide explains it better

[rapid kernel]

Hello everyone, during the mixing of two pth models, is it feasible to combine two index files also or what is the deal?

[rare goblet]

Ayo? @rapid kernel level 1 !!!

[proper shale]

Idk if you can even combine the index files tbh

You should definitely try the mixing models part tho, see what that gives

[rapid kernel]

yeah I will try mixing the index files into one file and try it with the new mixed model

[oblique venture]

How does the gender parameter in /speak in wordband bot work?

[delicate wharf]

where can I get someone to rate my models?

[hallow garnet]

Will this work with my microphone headset?

[untold dome]

Prediction failed.

index -1 is out of bounds for axis 0 with size 0

thoughts on how to fix?

[vagrant gale]

does anyone know how to add a model to applio rvc?

I have the pth and index file but im missing the resources tab

[royal panther]

RVC-GUI.bat isnt showing up on the extracted folder

[vagrant gale]

check antivirus

[lilac rose]

Yo what is the best girl voice changer for RVC?

[vagrant gale]

okada

[royal panther]

where is that? (sorry for asking)

[proper shale]

just add pth to the weights folder, index on logs

[proper shale]

just dont use rvc-gui

[vagrant gale]

will try thx

[proper shale]

use more recent local forks

-local

[azure marsh]

All Working Local Forks 🖥️

• 🍏 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

Need some help? 🤔

You can find more info on the "❔┋guides" channel. If you can't find your answer, feel free to ask for help in "help-rvc". Credits to Faze Masta and Antasma for compiling these links.

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

[proper shale]

maybe ur using trained index instead of added

[vagrant gale]

or just check windows defender or ur antivirus of choice

[royal panther]

oh my bad, then what else should I use

[proper shale]

stuff in here

[vagrant gale]

where's the weights folder located?

can't seem to find it

[proper shale]

its inside assets maybe?

[vagrant gale]

nope

just pycache , audios and i18n

[proper shale]

hm idk tbh, havent used appkio

applio*

[vagrant gale]

hm

[quiet ginkgo]

So I keep getting "Please select a model and input audio file" even if I selected and inputted both. Other models work but whenever I try to use the model I want to use, this error keeps popping up
RVC GUI btw

[vagrant gale]

someone helped me on another server, for applio rvc pth goes into logs/weights (which you have to create) and index into logs

[proper shale]

damn

[vagrant gale]

just posted this here if anyone will have the same problem as me in the future lol

[woeful citrus]

How many epochs should i use if im only using 9 seconds of audio? should i still use 100-200?

im reading the guide here and it says it varies based on your dataset, but mine is small, just 9 seconds so im not sure how many epochs use or if this will affect the quality of my AI

[brittle wing]

I assume it's not possible to convert a 40k sample rate model to a different sample rate? Outside of just re-training the model in the proper sample rate

Proceed with "Preprocessing", "Feature Extraction", and “Save preprocessed dataset files to Google Drive”.
does that mean run them in the gogle collab

@red kayak ?

sorry for pinging but your tags says helper

[uneven holly]

whats the best pitch_extraction_algorithm for a female voice model

[rare goblet]

Ayo? @uneven holly level 3 !!!

[half cove]

Try rmvpe

[uneven holly]

ok

how about crepe_hop_lengt

[half cove]

That won't affect rmvpe

[uneven holly]

ik

[half cove]

Crepe hop length fir mangio crepe

[brittle wing]

how do you fix this?

[rare goblet]

Ayo? @brittle wing level 6 !!!

[half cove]

I just use what it's set at for crepe

[half cove]

Did you run the dependency cell?

[brittle wing]

yes

[half cove]

Oh wait

I think

Did u set training veriables Before?

[brittle wing]

yes

[uneven holly]

save_frequency:50 is ok?

or 25

[red kayak]

I can help u in 30 mins

[molten pecan]

I recommend 25

[brittle wing]

nvmm i fixed it thx for offering ur help though

[molten pecan]

If you are a newbie

If you are, imo you should go for 20

[half cove]

What fixed it?

[brittle wing]

turns out i didnt run the triaing variables

i thought i just had to set them lol

[half cove]

I fucking knew it

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

https://docs.aihub.wtf/v/brazil

🇪🇸 Spain

https://docs.aihub.wtf/v/spain/

🇫🇷 France

https://docs.aihub.wtf/v/france

🇩🇪 Germany

https://docs.aihub.wtf/v/germany

🇯🇵 Japan

https://docs.aihub.wtf/v/japan

🇰🇷 Korea

https://docs.aihub.wtf/v/korea

🇵🇱 Poland

https://docs.aihub.wtf/v/poland

🇷🇺 Russia

https://docs.aihub.wtf/v/russia

🇮🇹 Italy

https://docs.aihub.wtf/v/italy

[half cove]

Cuz the cpu threads and experiment name are defined on the set training variables cell lol

[brittle wing]

lol

[brittle wing]

how many epochs should i do for 51 minute dataset?

[half cove]

Use tensorboard

[brittle wing]

how

[half cove]

It should open when you start training

To know how to use check out

-overtrain

[azure marsh]

How do I know if my model is overtaining?

All-In-One Guide on how to make a good model

This guide explains how the D and G files works and much more: https://rentry.org/RVC_making-models

Credits: LUSBERT

Automated Overtraining Detection (AOD)

Will be available soon in #1159513888199540817

Credits: grvyscale

[half cove]

In simple when the graph is closest to the bottom

[brittle wing]

how long should voice models be

datasets*

When I click run it does this

Is it supposed to be doing that @half cove ?

[uneven holly]

too late 50 goes brr

[vagrant gale]

what's the best pitch extraction algorithm?

tried using crepe but my pc froze

should I try harvest?

nvm im gonna use rmvpe

[orchid sapphire]

does rmvpe_onnx work just fine with nvidia 30 series cards or does it need to be changed to crepe_tiny
trying to help a friend set it up and i only know how to do it 100% with amd cards

[fathom ore]

can anyone send me a colab link to make covers?i have a model but the old colab stopped working

(im not trying to train a model i just need a colab to use it)

the guides here are confusing

[odd cape]

Is there a guide for how to use MahaTTS? I'm not sure how to add a file for the zero shot synthesis

[brisk salmon]

I tried to make a voice model with "Easy👨- 1/24/24" but now i can't find the added .index thing

is there a new update ? who changed soemthing ?

i have the pth but no the index anymore

[proper shale]

just recommend rvmpe

[proper shale]

check inside assets on the google drive, and then logs

or something like that

[proper shale]

-colab

[azure marsh]

Suggestions for @fathom ore

☁️ Google Colabs

• 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab

🤗 Hugginface 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
• GPT-SoVITS (demo), by litagin Huggingface Spaces

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

[brisk salmon]

i done and search .index , there nothing about an .index files

[rare goblet]

Ayo? @brisk salmon level 1 !!!

[proper shale]

then honestly i ain't sure

[brisk salmon]

i have only the pth files

[proper shale]

even on logs?

[brisk salmon]

yeah

[proven hill]

you can easily only train an index file afaik

takes 2-10 seconds

[brisk salmon]

oh ? where then

[proven hill]

(depends on the dataset but either way should not be too much)

[proven hill]

what are you using rn?

[brisk salmon]

this

[proven hill]

can you please send me the link?

[brisk salmon]

https://colab.research.google.com/drive/1r4IRL0UA7JEoZ0ZK8PKfMyTIBHKpyhcw

here

its in #1163883017446629508 message

easy gui thing

[little meteor]

guide for MDX-Net Colab please

[proper shale]

-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
• How to make a good model All-In-One guide, by LUSBERT Rentry
• Creating Datasets for RVC using iZotope RX, by LUSBERT Rentry
• Vocal Mixing Tutorial, by Roomie YouTube

🛠️ Tools

Youtube Downloaders

• https://github.com/yt-dlp/yt-dlp
• https://github.com/dsymbol/yt-dlp-gui

Audio Separation/Isolation

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

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

[proper shale]

https://docs.google.com/document/d/1HmkG9cmL8SLX7-vJcPT1-1KgUQtCrwXB8CicYmG4LW8/edit?usp=sharing explains it

[half cove]

Did you put experiment name as your set training variables

[half cove]

Did you train your model

[brittle wing]

how

[half cove]

The train model cell

[vapid gust]

stuck on 50% help

its a 10 second audio too loool

[brittle wing]

i clicked run and it did that

[half cove]

The model training one?

Send screenshot

[modest trout]

hii how i do this?

[brittle wing]

It’s stuck on this

[half cove]

What epoch are you at?

[brittle wing]

how to see

[half cove]

Scroll up on the cmd

[proper shale]

spaces in model name... that was prolly whh

[brittle wing]

so do i gotta restart

or

[brittle wing]

what should i change it to

[half cove]

Just remove space

For example

Your mom would be yourmom

[brittle wing]

ok

Is this good

[brittle wing]

IsaacClark(DS2-DS3).zip

[half cove]

That should work

[proper shale]

Folder inside dataset zip should havw no spaces too

[real vapor]

hello..

[brittle wing]

!!

aok

[real vapor]

do you know how to fix level low fatal error D3D on project playtime

[proper shale]

uhhhhh idk, this is a server related to ai so we can't reeeeally help you with that

[spare shale]

C:\Users\mikus>pip install customtkinter
Fatal error in launcher: Unable to create process using '"C:\Users\mikus\AppData\Local\Programs\Python\Python312\python.exe" "C:\Users\mikus\AppData\Local\Programs\Python\Python312\Scripts\pip.exe" install customtkinter': The system cannot find the file specified.

🙂

[brittle wing]

OK now i am on epoch 5

how long will it take?

[proper shale]

depends on the length of the dataset ngl

u should stop after 1 hour of training tho

[brittle wing]

ok

[brittle wing]

mine is 50 long minutes

[proper shale]

oh yeah it'll take a hot while

[brittle wing]

Is this good?

Also doesn’t collab kick you out for being afk

[proper shale]

yeeeah it's more than enough

[proper shale]

yeah, just check in every 5 mins

[brittle wing]

i will have to leave my house in an hour for like 4 hours

[brittle wing]

so is there any way to counter that

[proper shale]

so just stop right before that

and then u can continue later

[brittle wing]

but if i stop

would i kick me out anyways

[proper shale]

no

i mean yoy do need to export model n stuff

but uh

yeah thats about it

will kick you out after like 10-20 mins

[spare shale]

@proper shale How to fix this?

[proper shale]

yeah uh, there ain't no fixing that unless u have an NVIDIA GPU

[buoyant bone]

i cant hear myself

[brittle wing]

if i close the tab and reopen it some other time

will my progress still be there

[proper shale]

check https://rentry.co/W-Okada-FAQ

[proper shale]

if this is done quickly yes

byt uh

doesn't happen after a while

[spare shale]

@compact quest

[compact quest]

Hii

[rare goblet]

Ayo? @compact quest level 2 !!!

[brittle wing]

alright ill stop it now because i am going to take apart my pc to replace parts

[compact quest]

oh.

[spare shale]

You don't have it. Right?

[proper shale]

then export model

[brittle wing]

how

[proper shale]

stop training and run export model to drive step

[brittle wing]

alrihgt

Do I need to check anything off here?

[brittle wing]

wth is this error

[proper shale]

yeah you're fine

[proper shale]

no need

[brittle wing]

Ok

[rare goblet]

Ayo? @brittle wing level 7 !!!

[brittle wing]

How many epochs until it gets good

[proper shale]

Check this earlier version and see if it works (don't move any files out of it btw)

https://huggingface.co/wok000/vcclient000/resolve/main/MMVCServerSIO_win_onnxdirectML-cuda_v.1.5.3.15.zip

[proper shale]

It depends

Normally 50 and it already starts to sound ok

After 100 gradual improvements occur

[brittle wing]

So I should stop at 100?

[proper shale]

Nope

This varies a loooot

So keep training until you see it's overtraining or it stops improving in ur opinion

[brittle wing]

How do you see it overtraining

[ember coral]

whats the diffrent between onnx and pth file?

[proper shale]

TensorBoard

Also

-overtrain

[azure marsh]

How do I know if my model is overtaining?

All-In-One Guide on how to make a good model

This guide explains how the D and G files works and much more: https://rentry.org/RVC_making-models

Credits: LUSBERT

Automated Overtraining Detection (AOD)

Will be available soon in #1159513888199540817

Credits: grvyscale

[proper shale]

Guide explains how it all works

[proper shale]

pth is PyTorch (better for NVIDIA)
onnx is DirectML (better for AMD)

[ember coral]

oh okay thank you!

[proper shale]

you're welcome :)

[ember coral]

whats the best f0Detector within the voice changer?

[proper shale]

rvmpe

[north halo]

i forgot how to put the voices in

Like you put it in the folder and export it as onyx and whatever else

[north halo]

whats the folder called where you import the model pth and index?

or do you do it in the app

[proper shale]

pth in weights

index in logs

[stoic zealot]

for some reason, when I'm on a call with my friends using the ai voice changer, the ai listens to my friends' voices and repeats it in the model's voice, can someone explain why?

[proper shale]

check ur input/output settings in wokada and discord

make sure that w-okada's set up like this:

input - mic
output - CABLE Input

and that discord is like this:

input - CABLE Output
output - headphones

[north halo]

cant find them

[proper shale]

what fork r u using

[north halo]

idk what that means. cant I just click upload in the app and select it from my downloads?

[proper shale]

fork = variation of program

so im asking you basically what variation of rvc did u get

are you using w-okada?

[north halo]

the amd one

[north halo]

yeah

but I havent added a voice model in months so I forgot

[proper shale]

https://rentry.co/VoiceChangerGuide#uploading-models-amdintel-gpus-and-fixing-its-laggy-issue Here's the guide, you need to do some extra stuff so I sent this to make sure u prevent errors

[stoic zealot]

I put it the way you said, but it's still repeating my friends' voices

[proper shale]

u probably got some mic bleed or something... not sure what it is

[agile ledge]

Is this made by AI/ how was it made?

https://www.youtube.com/watch?v=JO47WRtw5tA

[proper shale]

it wasnt ai, definitely

dont know how youd recreate it tho

[foggy flame]

Hello guys does anybody knows how to make the voice from this tik tok? im googling it for 3h now and i didnt even come close because nobody even responded to comments on how its done. i tried to do with huggingface but it didnt work as inteded even in 1%

https://www.tiktok.com/@tipedxts/video/7327756161346850080

[agile ledge]

alr. thank you my good fellow.

[rare goblet]

Ayo? @agile ledge level 1 !!!

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

https://docs.aihub.wtf/v/brazil

🇪🇸 Spain

https://docs.aihub.wtf/v/spain/

🇫🇷 France

https://docs.aihub.wtf/v/france

🇩🇪 Germany

https://docs.aihub.wtf/v/germany

🇯🇵 Japan

https://docs.aihub.wtf/v/japan

🇰🇷 Korea

https://docs.aihub.wtf/v/korea

🇵🇱 Poland

https://docs.aihub.wtf/v/poland

🇷🇺 Russia

https://docs.aihub.wtf/v/russia

🇮🇹 Italy

https://docs.aihub.wtf/v/italy

[north halo]

their voices are probably coming out of your headphones loud enough for your mic to pick it up

[quasi flame]

Hi.

I have no idea how I ended up here.
I have no idea what this is.
I virtually have no idea about anything.

If I want to clone my own voice, but don't use one of these online services. Where do I start?

My goal:

• own my voice
• text to speech
• speech to speech

I already visited aihub.wtf but it appears to me that I already need to know what I do and need, to be able to understand the docs.

Can someone please push me in the right direction?

[upbeat sparrow]

whats the difference between inferencing and training?

(for ai voice acting)

[rare goblet]

Ayo? @upbeat sparrow level 1 !!!

[opal chasm]

does high epoch matter like is there a noticeable difference between 300 and 500 epoch models or na

[violet heron]

Inference is making cover

Training is making a model

[knotty parrot]

/colab

[rare goblet]

Ayo? @knotty parrot level 1 !!!

[violet heron]

-colab

[azure marsh]

Suggestions for @knotty parrot

☁️ Google Colabs

• 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab

🤗 Hugginface 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
• GPT-SoVITS (demo), by litagin Huggingface Spaces

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

[upbeat sparrow]

which is the best for model training?

[brittle wing]

I uploaded the Applio files and only edited the last few lines of the app.py to make it work

[upbeat sparrow]

is it better if all audio samples are in different files or all in one mp4?

[zinc nacelle]

The applio interface is so slow!

[dawn nebula]

没装python

你去装一下

@topaz plover

[topaz plover]

哪可以装

[dawn nebula]

等一下

你从哪里下载的

[topaz plover]

你说哪个

RVC?

[dawn nebula]

对啊

[topaz plover]

DC的

[dawn nebula]

你什么显卡

[topaz plover]

nvidia geforce gtx 1650

有人建议我下载原始

我下载原始的

[dawn nebula]

显存不够

搞colab吧

[topaz plover]

你说下载吗?

[dawn nebula]

colab是线上的

你想训练模型还是就推理

[topaz plover]

那个用了可以和朋友语音还是不能?

[icy chasm]

yo, got a ai software called "real-time voice changer client" and i was wondering how to setup like a output to microphone

i know i need to install something

any tutorials?

[dawn nebula]

可以

-realtime

[azure marsh]

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

[dawn nebula]

直接读就行，这个攻略写的还是比较直接明了的

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

https://docs.aihub.wtf/v/brazil

🇪🇸 Spain

https://docs.aihub.wtf/v/spain/

🇫🇷 France

https://docs.aihub.wtf/v/france

🇩🇪 Germany

https://docs.aihub.wtf/v/germany

🇯🇵 Japan

https://docs.aihub.wtf/v/japan

🇰🇷 Korea

https://docs.aihub.wtf/v/korea

🇵🇱 Poland

https://docs.aihub.wtf/v/poland

🇷🇺 Russia

https://docs.aihub.wtf/v/russia

🇮🇹 Italy

https://docs.aihub.wtf/v/italy

🎙️ Realtime Voice Changer (Local)

Guide for W-Okada's RealTimeVoiceChangerClient

Description: This one has more options and is somewhat more convenient to use than Real-time RVC.
Link: Rentry
Credits: ravencutie21

How to use real time conversion with RVC (voice changer)

Description: More accurate than W-Okada's, but requires a decent GPU.
Link: Google Docs
Credits: teas.hi

RVC TTS

Description: Text to speech with RVC
Link: Rentry

[topaz plover]

我开启MMV点击start_http.bat跑一下就没了

@dawn nebula

[dawn nebula]

有没有弹出来一个vc client

[topaz plover]

我点了有跑一瞬间进度，但就没了，也没有跑什么视窗

[icy roost]

can i get help some help involving my virtual cable here

[hot lagoon]

My trying to see how to run AICoverGen on paperspace keep geting the same thing "ModuleNotFoundError: No module named 'lxml.etree'

[timid canopy]

trying to download a model in google collab (rvc v2 with crepe), but i get this error for some reason:

ModuleNotFoundError: No module named 'mega'

i'm new to this stuff, so i'm not entirely sure what i'm doing-

[topaz plover]

@dawn nebula 在吗?

[sand iris]

mind sharing the link to it ?

as a workaround, replace the problematic thing with xml.etree instead of lxml.etree

Not sure how this one got through to begin with

[brittle wing]

hi, I want to save the model to drive (my training still continuining) which options should ı click for save to drive, I didnt udnerstand

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

https://docs.aihub.wtf/v/brazil

🇪🇸 Spain

https://docs.aihub.wtf/v/spain/

🇫🇷 France

https://docs.aihub.wtf/v/france

🇩🇪 Germany

https://docs.aihub.wtf/v/germany

🇯🇵 Japan

https://docs.aihub.wtf/v/japan

🇰🇷 Korea

https://docs.aihub.wtf/v/korea

🇵🇱 Poland

https://docs.aihub.wtf/v/poland

🇷🇺 Russia

https://docs.aihub.wtf/v/russia

🇮🇹 Italy

https://docs.aihub.wtf/v/italy

[low python]

does this mean that you now you need an Nvidia card as well for colab stuff!?

oh wait. it's doing something. It was static for a long time.

But in the end I will probably have done something wrong still. I am always fighting with these colab codes

I miss the easywebUI Gradio days. that shit was easy

[brittle wing]

what should I text here? If I didnt use the thing it says

what is g and d fiels

files

pls help

[forest sluice]

No

Never will be

Maybe go to Edit > Notebook Settings to see whether or not you're running a GPU?

[proper shale]

Nah it's fine, that message pops up when u don't use Fallback

If that's AICoverGen

[low python]

in the end it indeed worked itself out, but I wasn't able to make good covers with it. Probably user error. I am just used to the gradio that was shut down last summer by google. I haven't made covers since, so all the new methods only give me very shitty results cuz gradio was very noob friendly

[rare goblet]

Ayo? @low python level 2 !!!

[muted birch]

Hello, I want to make animation with artificial intelligence, can you help me?

[latent urchin]

I got a question. Do u even need the .index files in the AI voices? It takes so much space and when i enable it, it worsens the quality

[proper shale]

Index takes care of the accent, so uh, it is kinda needed

Unless you're using realtime

[proper shale]

But you can use the AI models without them either way

[brittle wing]

The extra list. Does it increase the sound quality of the model?

[proper shale]

Not really, it kinda makes the audio "smoother"

Makes the output better in a way

But increases delay and strains the CPU

[torn seal]

The latest collab in dev updates is giving me issues extracting the features. It is the one with the ui. I kinda need that

[pastel cove]

Hello

[proper shale]

I'd just not recommend using it because rvmpe uses CPU on that 😭

[torn seal]

It does?

That's gay

Not gpu

?

I guess if it gets the job done.

How do I get it to train the index then?

[proper shale]

...just use RVC disconnected

it's most likely gonna be a faster experience, without ui tho

[torn seal]

Yea I was thinking that

Plus the new one don't give you options like the disconnect

Can't choose rin e3

[proper shale]

yeah

[torn seal]

How is rin e3 ?

I havnt tried it? My guess is it doesn't need a big dataset?

[proper shale]

Haven't tested it too

Ov2 is that one that doesn't need a big dataset

[torn seal]

Oh

Oh really?

[gusty kestrel]

No results found for the search traviss...

Powered by Applio — Make sure you spelled it correctly!

[proper shale]

@prisma dune pls use #🔍│find-models

[torn seal]

I have a small dataset of a rapper named kadafi only 1min and 34 sec high quality.

🤔 💭

Ov2 huh , might try that

[proper shale]

Yea definitely try

[torn seal]

He's the guy on the song called me against the world with 2pac, it was his lil bro. Both dead 😢 😭

Luckily with this technology we can bring back the dead almost

[latent kettle]

Reduce batch size to 2

[brittle wing]

Where payday 2 voices

[proper shale]

Nah lol

4 would be enough ig

[proper shale]

#1175430844685484042

[fluid latch]

2 is too little

[brittle wing]

How do I get them fr fr

[fluid latch]

4 is perfect but if your dataset is dirty then 8

[latent kettle]

Just 1 minute 34 seconds ?? Do you think it is enough?

[brittle wing]

I just joined

[latent kettle]

Welcome!

[brittle wing]

All I need is payday 2 Bulldozer voice

[fluid latch]

yes you can get good models on datasets that are 1 minute long

[proper shale]

4 batch size yes i think it is

[fluid latch]

if the dataset is clean ofc

[proper shale]

he said the quality was great so

[latent kettle]

#🔍│find-models

[latent kettle]

Ohh ... I'm sorry for that!

[cyan juniper]

how to make ai sing?

[dawn nebula]

啥事

[dawn nebula]

你电脑配置不行跑colab吧，教程摆在那里自己看谢谢

1650 就4g，哪怕跑起来了占用率也会大，效果也不一定好

[upbeat sparrow]

is it more effective to separate audio clips into separate wavs or all audio clips should be into one long wav?

[proper shale]

Just keep em all in one wav

[proud thorn]

can someone help me for the voice changer

[proper shale]

What's up

Just explain the issue

[proud thorn]

nvm i find it thanks

[proper shale]

ah ok

[slim perch]

Anyone have some good resources for me to look at to learn about AI - I've played with it but I kinda want to do more than just play with image and text generators

[proper shale]

-rvc

[azure marsh]

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

❔ English FAQ

https://docs.aihub.wtf/faq

Credits: AI HUB FRANCE

✨ More guides

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

How To Make an AI Cover With Ilaria RVC

Link: Rentry
Credits: 👽 Julia (ailen2091)

[proper shale]

These should help

[proud thorn]

what the best model ia for girl

[proper shale]

idk

[turbid grail]

is #🔍│help-w-okada just dead now?

[vestal acorn]

what does this mean? this was in applio colab

[ember coral]

how can i make my dataset have higher quality with no backround noise and better and clearer vocals?

[rare goblet]

Ayo? @ember coral level 3 !!!

[weak palm]

get a better audio sample

[ember coral]

but there most be some app that can get it better

[weak palm]

Lets say you have a low quality youtube video

Do you think you can make that video higher quality

[ember coral]

i mean you could

[weak palm]

By changing to higher resolution ones youtube already has

You can remove background noise tho

with uvr

[vestal acorn]

what does this mean? this was in applio colab

[weak palm]

But it could make the audio worse

this is just with seperator and denoise

[ember coral]

oh okay but how do i use it i downloaded it

[weak palm]

You dont have to if you dont wanna spend time

Just send ur audio file here and ill run it through

If you want to

Otherwise yeah u gotta download it

[ember coral]

yeah im fine i can try doing it myself

[weak palm]

https://github.com/Anjok07/ultimatevocalremovergui/releases/tag/v5.6

[ember coral]

but are there anything you need to change in the program?

[weak palm]

just find the right ones

[rare goblet]

Ayo? @weak palm level 3 !!!

[weak palm]

dont know if theres better ones

this is for seperating vocals and music

this is for getting rid of annoying noise

and reverb HQ on the first one is to get rid of reverb

[ember coral]

so how do i download and use tensor board?

[white saffron]

-colab

[azure marsh]

☁️ Google Colabs

• 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab

🤗 Hugginface 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
• GPT-SoVITS (demo), by litagin Huggingface Spaces

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

[rare goblet]

Ayo? @white saffron level 2 !!!

[white saffron]

Guys, can anyone give me the Colab Ilaria RVC link?

[proper shale]

#📰│dev-updates

[half cove]

https://colab.research.google.com/drive/16LkwvFZeudTpUOsE_6bMjOq2qkxFo8Hr?usp=sharing

[white saffron]

Thanks

[ember coral]

how do i know if i have overtrained a model?

[royal panther]

Im new on windows, I dont know how to do files and stuff on here. This is going to sound really dumb but i remember watching the tutorial vid on RVC GUI and it told me to extract the folder. Is it supposed to be like this?

[dawn nebula]

-overtrain

[azure marsh]

Hey, @ember coral!

👇 Here are some resources to help you identify if your model is overtraining

How do I know if my model is overtaining?

All-In-One Guide on how to make a good model

This guide explains how the D and G files works and much more: https://rentry.org/RVC_making-models

Credits: LUSBERT

Automated Overtraining Detection (AOD)

Will be available soon in #1159513888199540817

Credits: grvyscale

[hidden dew]

hi, uh im trying to train a model in this notebook https://colab.research.google.com/drive/1r4IRL0UA7JEoZ0ZK8PKfMyTIBHKpyhcw#scrollTo=FFfC9x239kC1 and it says this error when i reach stage 4:

/usr/local/lib/python3.10/dist-packages/torch/utils/data/dataloader.py:557: UserWarning: This DataLoader will create 4 worker processes in total. Our suggested max number of worker in current system is 2, which is smaller than what this DataLoader is going to create. Please be aware that excessive worker creation might get DataLoader running slow or even freeze, lower the worker number to avoid potential slowness/freeze if necessary.
warnings.warn(create_warning_msg(
/usr/local/lib/python3.10/dist-packages/torch/nn/utils/weight_norm.py:30: UserWarning: torch.nn.utils.weight_norm is deprecated in favor of torch.nn.utils.parametrizations.weight_norm.
warnings.warn("torch.nn.utils.weight_norm is deprecated in favor of torch.nn.utils.parametrizations.weight_norm.")
DEBUG:infer.lib.infer_pack.models:gin_channels: 256, self.spk_embed_dim: 109
INFO:Anon:loaded pretrained assets/pretrained_v2/f0G40k_OV2.pth
Process Process-1:
Traceback (most recent call last):
File "/content/RVC/infer/modules/train/train.py", line 210, in run
utils.latest_checkpoint_path(hps.model_dir, "D*.pth"), net_d, optim_d
File "/content/RVC/infer/lib/train/utils.py", line 213, in latest_checkpoint_path
x = f_list[-1]

IndexError: list index out of range

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "/usr/lib/python3.10/multiprocessing/process.py", line 314, in _bootstrap
self.run()
File "/usr/lib/python3.10/multiprocessing/process.py", line 108, in run
self._target(*self._args, **self._kwargs)
File "/content/RVC/infer/modules/train/train.py", line 231, in run
torch.load(hps.pretrainG, map_location="cpu")["model"]
File "/usr/local/lib/python3.10/dist-packages/torch/serialization.py", line 986, in load
with _open_file_like(f, 'rb') as opened_file:
File "/usr/local/lib/python3.10/dist-packages/torch/serialization.py", line 435, in _open_file_like
return _open_file(name_or_buffer, mode)
File "/usr/local/lib/python3.10/dist-packages/torch/serialization.py", line 416, in init
super().init(open(name, mode))
FileNotFoundError: [Errno 2] No such file or directory: 'assets/pretrained_v2/f0G40k_OV2.pth'
训练结束, 您可查看控制台训练日志或实验文件夹下的train.log

[half cove]

Rvc gui is old

[dawn nebula]

holdup lemme read deez

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

https://docs.aihub.wtf/v/brazil

🇪🇸 Spain

https://docs.aihub.wtf/v/spain/

🇫🇷 France

https://docs.aihub.wtf/v/france

🇩🇪 Germany

https://docs.aihub.wtf/v/germany

🇯🇵 Japan

https://docs.aihub.wtf/v/japan

🇰🇷 Korea

https://docs.aihub.wtf/v/korea

🇵🇱 Poland

https://docs.aihub.wtf/v/poland

🇷🇺 Russia

https://docs.aihub.wtf/v/russia

🇮🇹 Italy

https://docs.aihub.wtf/v/italy

[dawn nebula]

have u runned everything correctly?

[hidden dew]

yes

i ran everything

top to bottom

[half cove]

I've only seen that error happen when inferring

[hidden dew]

oop

weird

[half cove]

What cell u running?

[dawn nebula]

try disable deez

[hidden dew]

both?

ok

[rare goblet]

Ayo? @hidden dew level 1 !!!

[dawn nebula]

how long is ur dataset

[hidden dew]

like seconds

[dawn nebula]

then just disable use ovr2

i need to test this notebook immediately

[brittle wing]

can someone watch me while ım doing the model from colab rvc disconnected, please its the only way I can do it

[hidden dew]

@dawn nebula

new error

[dawn nebula]

try start the whole thing again

i dont think it is installed correctly

[hidden dew]

oh

ok

[dawn nebula]

disconnect runtime

click runtime

and choose disconnect and delete runtime

[half cove]

I'm so confused rn

It uses

40k ov2 when 0v2 is selected

But

When its not selected

It uses 32k ov2?

Wildin

[hidden dew]

Im in the max 5 gc

vc*

streaming

[dawn nebula]

@half cove

i cant find 40kov2

[half cove]

Wait

Let me reread rq

[hidden dew]

Running this

@dawn nebula Is it possible that the error can be because it is an .ogg file

the dataset

[dawn nebula]

i havent use ogg as a dataset before

[dawn nebula]

try convert it to mp3 or wav

[hidden dew]

oh

ok

ill try

[half cove]

Check the code for training cell

It won't let me copy the code on mobile idk why

Fuck u colab

[hidden dew]

Nope wont work

[rare goblet]

Ayo? @hidden dew level 2 !!!

[hidden dew]

same error as the first i showed

[dawn nebula]

what the shit

[hidden dew]

Oop

[dawn nebula]

ok who wrote this notebook

[hidden dew]

is there another colab i should make

[half cove]

Told u

[hidden dew]

use*

if so link me to it

yea this one isnt working

[dawn nebula]

need a colab helper

[hidden dew]

whats that?

[half cove]

Rejekts I think

[dawn nebula]

a helper here who is far more knowledgeable in colab training

[half cove]

Just

[dawn nebula]

than me

[hidden dew]

oh

oop

[half cove]

Use disconnected

[dawn nebula]

-colab

[azure marsh]

☁️ Google Colabs

• 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab

🤗 Hugginface 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
• GPT-SoVITS (demo), by litagin Huggingface Spaces

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

[hidden dew]

wdym

[hidden dew]

o

ok ty

[half cove]

Rvc disconnected

[molten pecan]

@proper shale send guide for how to read tutorial

pleas

[half cove]

[dawn nebula]

💀

[brittle wing]

It says I need to run the Index Training cell AFTER model finishes but it is here for me BEFORE the training, So I didnt run it But it errors when I run it after training.

So I dont get the index file in my drive

[proper shale]

nah it ain't that bro

you actually need index before lol

it's probably the dataset formatting part u got wrong

[brittle wing]

Its the only one

before training

but it says I have to run it after training is finished

[proper shale]

Yeah that guide was written before the steps got rearranged

So yeah index should be before indeed

[brittle wing]

So

I have to cancel the training..

how do ı make it sotp

stop

without restrarting

everything

it always that eror when I try to run it ...

[proper shale]

yeah, because your dataset wasn't formatted properly

check the guide again, it explains how to format the dataset correctly

[brittle wing]

now I restart and run it before trainbing?

[proper shale]

nah you'll need to start from scratch it

cuz your dataset was formatted wrong

[brittle wing]

how

[proper shale]

literally, check the guide

[brittle wing]

do I format it I dont udnerstand anything omg

Theres nothing about it in the guide

[proper shale]

literally, the first thing talking about dataset

it explains you need to zip the folder with ur audios, with the same name as ur model

[brittle wing]

my dataset is already zio

[proper shale]

you might've screwed up on the folder name tho

[brittle wing]

its "guts"

ıll text everywhere it

everywhere I should type something

Im doing rn

Ill tell how it go

s

es

[proper shale]

Alright

[brittle wing]

I love u for helping my dumbass

"No features exist for this model yet. Did you run Feature Extraction?")

[half cove]

Ur not dumb dw

[brittle wing]

It says this but

I literally run the feature extraction

[brittle wing]

yes but Im lazy and my english is not the best

[half cove]

And you formatted it correctly?

[brittle wing]

ı guess so

what format means

I literally help a live help

I run it

I run the feature extraction correctly but ıdk why it says I didnt

:((

[dawn nebula]

no-feature-todo

[brittle wing]

what

[dawn nebula]

unfortunately u did not

[brittle wing]

how then

[dawn nebula]

dm me ur dataset

[royal panther]

hey so after I download all of this If i ever want to go back do I have to re-download it all over again?

[dawn nebula]

lol yeah

[royal panther]

yikes

[cobalt dome]

How to search for a model

[dawn nebula]

#🔍│find-models

then use / search function

[cobalt dome]

Ok

[molten pecan]

@half cove his pfp 😭

nah he'd win

which one of you would win though?

[cobalt dome]

I don't see it

The / command

[half cove]

https://media.discordapp.net/attachments/1159265300915687534/1181389516163002451/test.gif?ex=6580e1b9&is=656e6cb9&hm=f254a6f5a3267eea54fcc671df90f23558892ba98be1f1c11172b93e3bade625&

[half cove]

For?

[cobalt dome]

Searching

Nvm

[half cove]

Did you try in this channel or the search models channel

It should work

[cobalt dome]

Nvm I see it

[half cove]

[grizzled knoll]

how do i fix this error frequents occurred

[proper shale]

Check CMD

[brittle wing]

it gives me that error when I run the index training cell

[proper shale]

You most likely screwed up in the zip dataset part

[brittle wing]

I fckng didnt

Im tired 😩

what if I just skip it

[proper shale]

yesh either way, the model screwed up

because it didnt "feature extract" properly

[brittle wing]

HOW TO DO IT FCK

[proper shale]

probably try reexporting the dataset in audacity as one single wav

and then do the folder, zip thing

[outer willow]

hallu

hallou

[hollow dome]

Do any of y'all use Hina RVC primarialy? Do you know what to do about this error?
"'No model file exists in /content/Hina_RVC/rvc_models/[model name]"

[proper shale]

What Hina colab are you using

The AICoverGen one?

[hollow dome]

Yes

[proper shale]

Check if the download link for the model is working

And make sure you've put it in the right place too

[hollow dome]

How do I check?

[proper shale]

Open the link in another tab ig

[hollow dome]

What if its a model I'm uploading myself?

[proper shale]

Oh... I haven't really used that option so I'm not entirely sure

You can try uploading the zip to Huggingface and getting its link to download the model tho. I think that would solve it.

-huggingface

[azure marsh]

huggingface.co upload Guides

English guide

Link: https://rentry.org/fdg_guide_newer
Credits: FDG

Spanish Guide

Link: https://rentry.org/AI_Hub_guide_esp
Credits: Julia (ailen2091)

Hindi guide

Link: https://rentry.org/AI_hub_guide_hi
Credits: Enes

Turkish guide

Link: https://rentry.org/AI_hub_guide_tr
Credits: Enes

[proper shale]

Here's the tutorial btw

On how to upload to Huggingface

[hollow dome]

🫡

[proper shale]

Hope it works

[vestal isle]

/realtime

[rare goblet]

Ayo? @vestal isle level 1 !!!

[vestal isle]

`realtime

-realtime

[azure marsh]

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

[pearl tangle]

yo how do i use the voice models i download in the app

[rare goblet]

Ayo? @pearl tangle level 1 !!!

[rain magnet]

so uhm i want a vc changer do i jus download one of thee huggin links

[rare goblet]

Ayo? @rain magnet level 1 !!!

[rain magnet]

im not sure how i do this shi

[brittle wing]

It would be helpful to send images so I can explain this better but the voice I put it won't show up on Mangio

I try using the download thing and it still doesn't work

[sand iris]

You should get image perms after sending a few messages

[brittle wing]

Damn

I even tried the default AI it gives me and still nothing]

[rare goblet]

Ayo? @brittle wing level 1 !!!

[brittle wing]

Oh there we go

The arrow doesn't work either

[peak hamlet]

ye the colab thing seems to be shitting itself a little bit

[normal topaz]

crepe isnt showing up for me

[peak hamlet]

I also get this error when I try to load any model

https://imgur.com/a/ATJ1lBl

[brittle wing]

Oh yeah now I notice

[sick warren]

does somebody knows the settings for Iris Heart (Neptunia) (RMVPE, Ov2) (RVC v2, 200 epochs)?

[open cradle]

i keep getting interaction failed when I try to download a model from here

[sick warren]

i dont think we get help either

[rare goblet]

Ayo? @sick warren level 1 !!!

[open cradle]

:/

[brittle wing]

how can i make models

[proper shale]

#1159513888199540817

[proper shale]

Try using another bot to search

Or just use #1175430844685484042

[open cradle]

can you elaborate? which bot should I use and how do I use it

[rare goblet]

Ayo? @open cradle level 1 !!!

[open cradle]

it doesnt have the one that im looking for

i was looking for travis scott

[proper shale]

There are two Travis models in #1175430844685484042 , i think you can search there

[proper shale]

Maybe try the Applio bot? Or ModelAI?

Not sure, I don't really use voice models tab a lot

[violet saffron]

MJ, você pode me ajudar?

[outer willow]

heyllow

[olive cliff]

Hi. I have a question. I'm currently training a model on the RVC v2 Disconnected Colab notebook. I just wanted to ask if I were to resume training, do I have to run "set training variables" before loading the preprocessed dataset files and importing the model?

[proper shale]

You do need to run Set Training Variables and insert the same settings as last time.

[night dove]

High graphics set fps lock over 60, Ideally 120, turn direct x 12 on

[dawn nebula]

hi, does anyone know how to manually edit f0 for inference

[olive cliff]

Okay. Thanks.

[misty stream]

how do i make rvc not open my browser every time i start it

[proper shale]

You're welcome :)

[proper shale]

I think this is something inside the code tbh, not something you can change.

...if you know coding though, maybe

maybe

[misty stream]

i know how to comment lines and make things stoip working this seems like it should be within my ability as a programmer

[balmy citrus]

Is it easy to improve AI models with more audios?

[proper shale]

Yeah, if it's your model

Then u just gotta train a new model and see how it sounds

[balmy citrus]

i mean can i put just new audios in the ai model or does it have to be a new AI model?

[proper shale]

Yeah it has to be a new one, else it'll screw up on training

[balmy citrus]

okay thank you

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

https://docs.aihub.wtf/v/brazil

🇪🇸 Spain

https://docs.aihub.wtf/v/spain/

🇫🇷 France

https://docs.aihub.wtf/v/france

🇩🇪 Germany

https://docs.aihub.wtf/v/germany

🇯🇵 Japan

https://docs.aihub.wtf/v/japan

🇰🇷 Korea

https://docs.aihub.wtf/v/korea

🇵🇱 Poland

https://docs.aihub.wtf/v/poland

🇷🇺 Russia

https://docs.aihub.wtf/v/russia

🇮🇹 Italy

https://docs.aihub.wtf/v/italy

[proper shale]

You're welcome :) Good luck on the model!!

[narrow compass]

how do i upload a model to llara

[proper shale]

Download Voice Models section on the UI

u need a Huggingface link

[narrow compass]

yeah but then it says link invalid

[rare goblet]

Ayo? @narrow compass level 1 !!!

[narrow compass]

https://huggingface.co/orhay1/RVC_Belle_EN/tree/main

[proper shale]

What link were u using

[proper shale]

Oh

[narrow compass]

trying to use belle singing voice

[proper shale]

You need to copy the link of the zip download button

[narrow compass]

"You need to name your model. For example: My-Model"

[proper shale]

Right click on the download button of the zip file, then copy link and paste there

[narrow compass]

but when i do that it says download invalid

[proper shale]

Yeah just name it anything

[narrow compass]

yeah i did

"ERROR - The download failed. Check if the link is valid."

[proper shale]

https://huggingface.co/orhay1/RVC_Belle_EN/resolve/main/Belle_Singing.zip

this

and then just name ur model anything

[narrow compass]

thx

[proper shale]

yw :)

[molten pecan]

Menh and mrm0dz gave the thumbs up for sharing the site

It's around the corner!

[half cove]

[proper shale]

Wooo

Putting on my clipboard definitely

[rare goblet]

Ayo? @brazen merlin level 1 !!!

[hexed garnet]

HOLA

alguien me ayuda

a configurar RVC w./ w-okada

[tranquil cipher]

how do you get rid of pipline is not initalized

[vestal crow]

Where can I access Turkish voice models?

[dawn nebula]

#🔍│find-models

[vestal acorn]

what does "failed to train index: need at least 1 array to concatenate" mean? (this was made in applio colab)

[small scarab]

Dumb question but if I want to use a custom pretrain in Applio 3.0.4 do I have to tick off "Pretrained" and only keep "Custom Pretrained" on?

[rare goblet]

Ayo? @small scarab level 1 !!!

[small scarab]

It's a bit confusing for me since the console doesn't really say which pretrain it is using.

[dawn nebula]

send ss

ur family guy models are great btw

[small scarab]

This is what I've set for training and I'm using the RIN_E custom pretrain. Also thank you lol

Yeah I'm not too sure if I'm supposed to keep both of them ticked or just keep custom pretrained on and not pretrain

A lot has changed and I'm trying to figure out how everything works again lol

Aside from that I believe I've got everything else under control

[dawn nebula]

ask applio devs ig i use mainline now

[wheat plover]

I haven't been here in a while, what's the current best way to train models online on google colab?

[rare goblet]

Ayo? @wheat plover level 3 !!!

[wheat plover]

or like, the most up to date tutorial

[violet heron]

Which colab?

[wheat plover]

any

just need something good

[violet heron]

Here’s the one for the RVC Disconnected colab https://docs.google.com/document/d/1XuxQYiqEhYrdYeCZRRLrmV_ciMKo0bV-jTCGHu_-5Cc/edit?usp=drivesdk

[wheat plover]

thank you

[limber forge]

Heyo

So just got the voice changer client and it seems I can only use 2 of the F0 det features. The only ones I can use is crepe tiny and rmvpe onnx. Is there anyway I can get to use the 5 other options? They have (N/A) next to the parts I can't use.

[violet heron]

What’s your GPU?

[limber forge]

Nvidia Geforce RTX 4060Ti

[rare goblet]

Ayo? @limber forge level 1 !!!

[violet heron]

Does it say DirectML

[limber forge]

In task manager it doesn't say DirectML

[violet heron]

No on Okada

[limber forge]

Oh

[violet heron]

The “Real Time Voice Changer Client Demo”

Does it have DirectML next to it

[limber forge]

Are you talking about the top right area with small text? says onnxdirectML-cuda

[violet heron]

You downloaded the wrong version

[brave garnet]

Voicechanger Download (Okada)

⠀
Download for Nvidia GPUs
Version 17b cuda

Download for AMD GPUs
Version 17b directml

Download for Intel GPUs
Version 17b directml

Download for Mac
Version 17b Mac
⠀

Made with love by Ilaria 💖

[limber forge]

Aaahhhggg

Thanks -w-

Guessing I should download the Nvidia one?

[violet heron]

Yes

[karmic moat]

Traceback (most recent call last):
  File "E:\AI\Applio-3.0.4\Applio-3.0.4\env\lib\multiprocessing\process.py", line 315, in _bootstrap
    self.run()
  File "E:\AI\Applio-3.0.4\Applio-3.0.4\env\lib\multiprocessing\process.py", line 108, in run
    self._target(*self._args, **self._kwargs)
  File "E:\AI\Applio-3.0.4\Applio-3.0.4\rvc\train\train.py", line 259, in run
    train_and_evaluate(
  File "E:\AI\Applio-3.0.4\Applio-3.0.4\rvc\train\train.py", line 441, in train_and_evaluate
    wave = commons.slice_segments(
  File "E:\AI\Applio-3.0.4\Applio-3.0.4\rvc\lib\infer_pack\commons.py", line 49, in slice_segments
    ret[i] = x[i, :, idx_str:idx_end]
RuntimeError: The expanded size of the tensor (12800) must match the existing size (12600) at non-singleton dimension 1.  Target sizes: [1, 12800].  Tensor sizes: [12600]
Index training finished!```

applio 3.0

takes a pretty long time to train just a couple epochs and then randomly stops training with this error

completely resetting all progress after the last checkpoint

just pressing train again immediately afterwards works fine until it breaks again

[slate mortar]

how do i get the voice changer to work in games

downloaded the vb virtual thing

and output is in the input and the input is in the output for some reason

[azure marsh]

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

https://docs.aihub.wtf/v/brazil

🇪🇸 Spain

https://docs.aihub.wtf/v/spain/

🇫🇷 France

https://docs.aihub.wtf/v/france

🇩🇪 Germany

https://docs.aihub.wtf/v/germany

🇯🇵 Japan

https://docs.aihub.wtf/v/japan

🇰🇷 Korea

https://docs.aihub.wtf/v/korea

🇵🇱 Poland

https://docs.aihub.wtf/v/poland

🇷🇺 Russia

https://docs.aihub.wtf/v/russia

🇮🇹 Italy

https://docs.aihub.wtf/v/italy

[empty blade]

guys is there a dream the minecraft youtuber r vc

rvc

I can't find it anywhere

[brittle wing]

anyone know how to fix this

[outer willow]

heyllow

heyllow

[proper mist]

Hello! Starting to use AI voice conversion for the first time, I'm wondering how to use the voice models here, not looking for real-time conversion, just want to convert some wav files using a voice model.

[tame mica]

-rvc

[azure marsh]

Documentation

📚 All-In-One English documentation

https://docs.aihub.wtf/

❔ English FAQ

https://docs.aihub.wtf/faq

Credits: AI HUB FRANCE

✨ More guides

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

How To Make an AI Cover With Ilaria RVC

Link: Rentry
Credits: 👽 Julia (ailen2091)

[tame mica]

and -colab

damn

-colab

[azure marsh]

☁️ Google Colabs

• 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab

🤗 Hugginface 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
• GPT-SoVITS (demo), by litagin Huggingface Spaces

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

[tame mica]

^

[proper mist]

Thank you!

[modern surge]

-colab

[azure marsh]

☁️ Google Colabs

• 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab

🤗 Hugginface 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
• GPT-SoVITS (demo), by litagin Huggingface Spaces

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

[stiff goblet]

Is 30 minutes dataset too long for ov2 pretrain?

[proper mist]

I managed to convert some audio and the previews sound great but extracting the audio is a problem for me. Whether I set an output folder to automatically extract the audio files or just download them individually, the downloaded quality is horrible and they sound nothing like the previews.

[upbeat sparrow]

i cant connect to t4 in mdx-net collab

[rare goblet]

Ayo? @proper mist level 1 !!!

[sweet leaf]

how do i use this is there i sight i need to go to?

[quaint hound]

what's the actual best way to laugh w/ the voice changer without it sounding weird

[brittle wing]

question: If i use an american model but use a different index file from a german person. Would it solve the accent problem?

[gray crow]

-colab

[azure marsh]

☁️ Google Colabs

• 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
• RVC Disconnected, by Kit Lemonfoot Google Colab
• easyGUI, by rejects Google Colab

🤗 Hugginface 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
• GPT-SoVITS (demo), by litagin Huggingface Spaces

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

RVC Guides (How to Make AI Cover)

Documentation

🇺🇸 English (main)

https://docs.aihub.wtf/

​

Translation by country

🇧🇷 Brasil (PT-BR)

https://docs.aihub.wtf/v/brazil

🇪🇸 Spain

https://docs.aihub.wtf/v/spain/

🇫🇷 France

https://docs.aihub.wtf/v/france

🇩🇪 Germany

https://docs.aihub.wtf/v/germany

🇯🇵 Japan

https://docs.aihub.wtf/v/japan

🇰🇷 Korea

https://docs.aihub.wtf/v/korea

🇵🇱 Poland

https://docs.aihub.wtf/v/poland

🇷🇺 Russia

https://docs.aihub.wtf/v/russia

🇮🇹 Italy

https://docs.aihub.wtf/v/italy

[whole briar]

Hello, does someone know if there is a well explained way to have tts on twitch (for example redeeming reward points) with my own trained rvc models?

[sick warren]

how about all settings like index, gain, tune, chunk or extra?

[wise sigil]

can someone dm me i need help 😭

what this even mean

[proper shale]

delete model_dir and restart W-Okada

[wise sigil]

is that in the folder

[proper shale]

yes

[wise sigil]

how about model_dir_static

im trying now

[wise sigil]

i dont know why but i dont have any output from the program

makes no sence to me honestly i followed to tutorial

im on a laptop i know but i dont hear anyy sound at all

[proper shale]

wait do you even have a GPU