Tyy
#✨│ai-help
viral mason
good luck 
viral mason
I know how to use kaggle and wokada tg but not the browser version like this
viral mason
Virtual cable again
ye
graceful furnace
Aight
viral mason
actually no not if u have it already
graceful furnace
Ohh oke
So this means when i wanna vc with the voice changer i gotta use discord ob my browser as well
upper juniper
the Intel ARC GPUs also bad? I mean a CPU/APU I can get, that is bottom of the performance barrel, but the dedicated GPUs also?
graceful furnace
Oh yea you did say thay mb😭
viral mason
the Intel ARC GPUs also bad? I mean a CPU/APU I can get, that is bottom of the p...
dedicated gpus just cannot run ai at all
not locally
upper juniper
then what do they even have all that VRAM for xD
viral mason
then what do they even have all that VRAM for xD
honestly no clue lol
graceful furnace
Konichiwa
young furnace
i mean i will put my own ai models, i know how but generally spongebob mario v1 from ultrakill etc.
mainly from games and shows
viral mason
i mean i will put my own ai models, i know how but generally spongebob mario v1 ...
wonderful! was checking because a lot of ppl join here for weird stuff
here's the downloads!
first the voice changer second is a virtual audio cable
young furnace
wonderful! was checking because a lot of ppl join here for weird stuff
really? bruh
i just like to roleplay my favorite characters
so dw
and thanks for software!
viral mason
and thanks for software!
no problem! if u have questions lemme know
viral mason
really? bruh
yea, ppl can be weird lol
rare tide
Does anyone who do models take suggestions? My pc isn't good for training so :/
copper wraith
Wasnt there a server network for finding voice models a while back? I remember there was a seperate server where you could get german voice models but i cant find it, does anyone have the link?
uneven plank
When I click a huggface link to download a voice model I got: Invalid username or password.
viral mason
Does anyone who do models take suggestions? My pc isn't good for training so :/
of course! I can send a video I've made a tutorial video not on yt but one I can send here
viral mason
https://huggingface.co/KJAV/RVC/resolve/main/HitoriGotoh.zip
what are you trying to download?
errant perch
- Goal: TTS
- Specific Issue: doesn't work
- Full GPU Name: MSI Geforce rtx 5070ti 16gb
- Operating System: win 11
- Tutorial Link used:
viral mason
what are you using for tts?
viral mason
could you show me the link you downloaded it from? Applio is the best tts if you're using voices from here https://discord.com/channels/1159260121998827560/1175430844685484042
errant perch
errant perch
oh yeah real time i think
viral mason
real time is for talking into it and it makes your voice a different one, that's what you're wanting right?
like a voice changer
viral mason
it's alright
since you have a 5070ti I will get you Vonovox, it's the current best for Nvidia
the first link is the voice changer
the second link is a virtual audio cable to connect the voice changer to discord or games
<@&1159293140440723499> hacked account
if I had my mod status back I could deal with this. ugh
wide oasis
+yo
fathom rapids
Hey guys. How everythign is going?
Need help seems the this set up is no longer working.
MMVCServerSIO
errant perch
https://huggingface.co/dr87/vonovox/resolve/c8034f5f6d50648a8109bb4f847182362e2b...
is it better then the one i have rn?
unborn umbra
when the cmd pop up i dont understand how i can open the voice changer
viral mason
is it better then the one i have rn?
massively better, vonovox is the best currently for realtime as it has improved quality and is much more optimized so it will run better
viral mason
Hey guys. How everythign is going?
Need help seems the this set up is no longer ...
what gpu do you have (Nvidia or AMD) and what are you using it for?
viral mason
when the cmd pop up i dont understand how i can open the voice changer
what gpu do you have (Nvidia or AMD) and what are you using it for?
fathom rapids
what gpu do you have (Nvidia or AMD) and what are you using it for?
GPU: NVIDIA GeForce RTX3050.
Im using it for voice dubbing for shorts clips 😄
unborn umbra
what gpu do you have (Nvidia or AMD) and what are you using it for?
i did it
viral mason
GPU: NVIDIA GeForce RTX3050.
Im using it for voice dubbing for shorts clips 😄
cool! u should try out Vonovox, it's currently the best
https://huggingface.co/dr87/vonovox/resolve/c8034f5f6d50648a8109bb4f847182362e2b779b/Vonovox_beta_17_11.zip
fathom rapids
cool! u should try out Vonovox, it's currently the best
https://huggingface.co/d...
Do i need to delete the files that I have? and DL vonobox?
viral mason
it would be good to delete it in case the old one messes with Vonovox for whatever reason
u can keep the models u were using tho
errant perch
what is that
fathom rapids
cool! u should try out Vonovox, it's currently the best
https://huggingface.co/d...
by any chance do you have a walkthrough vid how to use it(connecting to other apps and uploading models)?
viral mason
what is that
it's the second link I sent
errant perch
i already have virtual cable
viral mason
by any chance do you have a walkthrough vid how to use it(connecting to other ap...
what part are you confused on, just making sure so I can show you the right part
rare tide
of course! I can send a video I've made a tutorial video not on yt but one I can...
It's just I have an amd gpu so
fathom rapids
what part are you confused on, just making sure so I can show you the right part
ill run vonovox first then ill get back to you 😄
viral mason
alrighty
viral mason
That voice model
did you download it from here? https://discord.com/channels/1159260121998827560/1175430844685484042
uneven plank
yes
viral mason
clicking the link shows it's not working
that means it was most likely deleted, nothing can be done
uneven plank
F
viral mason
you could always remake the model yourself, I made a guide in the vonovox server that shows how to get the voice you want and how to train it
rare tide
I just want a good meiko and kaito, adachi rei and like narita taishin 😔
viral mason
i dont have start.bat on my file. Which one will I use to run vonovox?
wdym you don't have start.bat? try redownloading it, that file should come with it when installing
viral mason
It's just I have an amd gpu so
it should be fine to run on Kaggle, which is in browser
follow this
slim linden
- Goal: realtime
- Specific Issue: not sure this is the right spot to ask, but i downloaded vonovox and tested a model, and now when trying to move/rename the folder it's in it doenst and it says 'the folder or a file in it is open in another program'. not sure if this is a usual problem but i cant find anything in my task manager that's holding this up (after closing vonovox and everything i could find)
- Full GPU Name: RTX 4060 ti
- Operating System: windows 10
viral mason
- Goal: realtime
- Specific Issue: not sure this is the right spot to ask, but i...
tts? vonovox is for real time, like talking into it and it saying the same with whatever ai voice youu chose
slim linden
sorry right i meant real time 
my bad
fathom rapids
wdym you don't have start.bat? try redownloading it, that file should come with ...
i think its now running. how can i monitor the voice?
viral mason
sorry right i meant real time <:boink:1451113890556411985> my bad
all good, you can rename a voice in vonovox by right clicking it I believe inside the app
cyan kettle
all good, you can rename a voice in vonovox by right clicking it I believe insid...
Where are the sounds located?
models?
viral mason
i think its now running. how can i monitor the voice?
in a moment I can show you, but go to control panel > voice settings > recording > and enable listen to this device on the virtual audio cable
slim linden
all good, you can rename a voice in vonovox by right clicking it I believe insid...
i mean the overall folder where vonovox is stored. im more wondering because it would mean there's still a file running in my system from vonovox thats hidden somewhere in my programs (or so i think). not sure if its been a general problem or not. maybe its just a one time predicament ive gotten
viral mason
i mean the overall folder where vonovox is stored. im more wondering because it ...
I'm unsure tbh, I think it would be best to ask in the official vonovox server
slim linden
alright thank you
cyan kettle
https://discord.com/channels/1159260121998827560/1175430844685484042
I need a link to a website that contains audio files so I can search for them on Google.
viral mason
I need a link to a website that contains audio files so I can search for them on...
what are you looking for?
viral mason
alright thank you
you're very welcome!
cyan kettle
what are you looking for?
The best voices in terms of quality and performance
viral mason
https://discord.com/channels/1159260121998827560/1175430844685484042
if you want voices they're here
are you confused?
cyan kettle
What is the difference and which is better?
fathom rapids
Im all good now. Thanks @viral mason
prisma grove
quick question, does it matter whether I train a voice model on applio or mainline rvc
like, is it gonna sound better? train faster?
also, is using a pretrain a viable option for creating a voice model of a specific character? or is it just going to make it sound like a random person instead?
wraith oar
how do i install the program?
hardy yew
quick question, does it matter whether I train a voice model on applio or mainli...
The only differences are the settings and features, but training code is shared anyway. My vote is for Applio
hardy yew
also, is using a pretrain a viable option for creating a voice model of a specif...
You always want to use a pretrain. You can't train a particular voice from scratch without one. It won't work like that
The pretrain contains a lot of knowledge, including how to generate the spectrogram at all. Then fine-tuning it for the particular voice is what makes it just learn to sound like it.
glacial bear
any good local RVC applications suggestions for macbook air m4 ???
hallow thistle
any good local RVC applications suggestions for macbook air m4 ???
Retrieval-based voice conversion or realtime voice changer? And what will you use the software for?
glacial bear
Retrieval-based voice conversion or realtime voice changer? And what will you us...
a realtime voice changer and im planning to show it to a friend to spread awareness on how dangerous it could be
i think a retrival based one would work but im not sure if i can send the clips so i would prefer a realtime one
hallow thistle
glacial bear
<:Mahiro_stares:1462023303609843816>
????
hallow thistle
There's Tg Develop's voice changer for Mac Silicon. Follow the guide https://docs.aihub.gg/realtime-voice-changer/local/tg-develops-w-okada-fork/#mac, and download https://github.com/tg-develop/voice-changer/releases/download/b2364/voice-changer-macos-arm64-cpu.tar.gz.
glacial bear
thanks a lottt
sinful depot
guys is this server purely for rvc? or can i ask about other ai too
hardy yew
guys is this server purely for rvc? or can i ask about other ai too
it's general, so feel free to ask
prisma grove
The pretrain contains a lot of knowledge, including how to generate the spectrog...
how do you fine tune it?
hardy yew
just train a model using a pretrain
that's effectively finetuning the pretrain for a particular voice
this is pretty much how all models in #1175430844685484042 are made
prisma grove
how do I know which one to use?
low shard
any good local RVC applications suggestions for macbook air m4 ???
!help-template
patent trellisBOT
!help-template
low shard
guys is this server purely for rvc? or can i ask about other ai too
it was purely for RVC, now its general ai
upbeat imp
what to use instead of weight ai for covers
hallow thistle
what to use instead of weight ai for covers
What is your PC GPU? There's Applio RVC.
upbeat imp
gtx 960
lyric fossil
- Goal (e.g., TTS, AI Covers, Roleplay): AI voice conversion cloud using Kaggle
- Specific Issue: There's a output error in the log ```2026-05-17 13:36:41.721330963 [E:onnxruntime:Default, cudnn_fe_call.cc:33 CudaErrString<cudnn_frontend::error_object>] No valid engine configs for ConvFwd_
2026-05-17 13:36:41.722017527 [E:onnxruntime:, sequential_executor.cc:516 ExecuteKernel] Non-zero status code returned while running Conv node. Name:'/hubert/encoder/pos_conv_embed/conv/Conv' Status Message: Failed to initialize CUDNN Frontend/onnxruntime_src/onnxruntime/core/providers/cuda/cudnn_fe_call.cc:99 std::conditional_t<THRW, void, onnxruntime::common::Status> onnxruntime::CudaCall(ERRTYPE, const char*, const char*, SUCCTYPE, const char*, const char*, int) [with ERRTYPE = cudnn_frontend::error_object; bool THRW = true; SUCCTYPE = cudnn_frontend::error_code_t; std::conditional_t<THRW, void, common::Status> = void] /onnxruntime_src/onnxruntime/core/providers/cuda/cudnn_fe_call.cc:91 std::conditional_t<THRW, void, onnxruntime::common::Status> onnxruntime::CudaCall(ERRTYPE, const char*, const char*, SUCCTYPE, const char*, const char*, int) [with ERRTYPE = cudnn_frontend::error_object; bool THRW = true; SUCCTYPE = cudnn_frontend::error_code_t; std::conditional_t<THRW, void, common::Status> = void] CUDNN_FE failure 8: HEURISTIC_QUERY_FAILED ; GPU=1 ; hostname=c2e0cb52600b ; file=/onnxruntime_src/onnxruntime/core/providers/cuda/nn/conv.cc ; line=225 ; expr=s_.cudnn_fe_graph->create_execution_plans({heur_mode});
- Full GPU Name: GPU T4 X2
- Operating System: Windows 10
- Tutorial Link used: https://docs.aihub.gg/realtime-voice-changer/cloud/tg-develops-w-okada-fork-cloud/
- Exact time when it happened : after i click the model to start the ai voice conversion (on ngrok GUI)
still gull
hello
anyone have experience dealing with enterprise AI plan , i need advice on which option is better for enterprises using AI , a custom wrapper that have DLP( we need this) or enterprise plan purchase is a better option in any case
outer isle
Is the weights.com working? because I can't access to it?
hallow thistle
Is the weights.com working? because I can't access to it?
Have you missed the news? The Weights.com has shut down few months ago.
outer isle
Yes, but i Can't access to previous version of Replay?
lyric fossil
- Goal (e.g., TTS, AI Covers, Roleplay): AI voice conversion cloud using Kaggle...
anyone know why this happened?
proven hill
- Goal (e.g., TTS, AI Covers, Roleplay): AI voice conversion cloud using Kaggle...
@viral mason
i know you use kaggle
viral mason
- Goal (e.g., TTS, AI Covers, Roleplay): AI voice conversion cloud using Kaggle...
Yo! I use Kaggle, are you trying to use it for simple conversion like ai cover or realtime talking?
oh
nvm
lyric fossil
i've used it before, like 3 months ago, never got this issue
viral mason
have you looked at the guide to see if it may explain any issue you may have?
I know nothing at all on how to run the voice changers on kaggle just applio
outer isle
Yes, but I can't access to the previous version of Replay
lyric fossil
have you looked at the guide to see if it may explain any issue you may have?
which one is the troubleshoot guide for realtime ai voice kaggle?
outer isle
@hallow thistle
outer isle
When I go to Weights.com and I get the “Page Not Found” message. I need to access the previous version of replay. I’m on Windows 10
graceful furnace
How do I make this sound less fucky </3
I used replay and I didn’t touch anything else
@viral mason
proven hill
why yall use replay its garbage
outer isle
When I go to Weights.com and I get the “Page Not Found” message. I need to access the previous version of replay. I’m on Windows 10 with a Intel Core 13th generation
hallow thistle
why yall use replay its garbage
Sorry if I sound mean, but I just think those who go for Replay mostly prefer the easy-to-use than having many features like in Applio RVC, and it's kind of a bit understandable at least to me.
outer isle
I’m trying to transfer the Song Ai cover and when I go to Weights.com and get the latest version of Replay, I get the page not found message
Has anybody has the problem?
graceful furnace
why yall use replay its garbage
Thats what im used to using!!
What do you use?
viral mason
Thats what im used to using!!
What do you use?
replay is ok but there's also ai cover maker
graceful furnace
replay is ok but there's also ai cover maker
Whats it calleddd
viral mason
Whats it calleddd
ai cover maker 😭
graceful furnace
Hold on pleas
hallow thistle
AI Cover Maker (or AICoverMaker) is basically another RVC fork.
graceful furnace
https://docs.aihub.gg/rvc/local/aicovermaker
Im bsck
Tyy
viral mason
you're welcome! I haven't tried this but I heard Nick say it's like replay
west temple
im getting 190 perf, using 196ms only with 2.7 sec
final elk
Yo! I use Kaggle, are you trying to use it for simple conversion like ai cover o...
Teach me for a.i cover
grave shard
viral mason
Teach me for a.i cover
the easiest option is to use either ai cover maker or replay but Applio is also an option tho that one requires you to seperate the vocals and music for it to work properly
karmic trellis
Hi i have been using vonovox and wokada for a few months now, but i want to begin learning how to make/train my own models, is there a tutorial video or some sort of a guide for that somewhere online?
viral mason
Hi i have been using vonovox and wokada for a few months now, but i want to begi...
Hi there! I made a super easy to understand 2 part tutorial for the Kaggle version of Applio, kaggle is like google colab but much better as it gives 30 hours a week for free users
applio is the main training software used for model making, lemme know if you're interested
karmic trellis
Hi there! I made a super easy to understand 2 part tutorial for the Kaggle versi...
can you send me a link for that tutorial?
viral mason
can you send me a link for that tutorial?
final elk
the easiest option is to use either ai cover maker or replay but Applio is also ...
I used these but error
viral mason
I used these but error
if you follow the kaggle tutorial as shown there should not be any errors
final elk
if you follow the kaggle tutorial as shown there should not be any errors
Can you send it
dense topaz
Do i ask here for help?
viral mason
Do i ask here for help?
yep
shut dagger
Hello, the bot told me to post and here, and I want to ask for anybody who a model maker role can you review my model and if you like it upvote it? It's a Lil Baby model
jagged blade
how do u make voice models for free? is it possible
- Goal making a voice
- Specific Issue: idk how gng
- Full GPU Name: AMD Ryzen 9 9900X processor, 64GB DDR5 RAM, a fast 2TB NVMe SSD plus 4TB additional storage, powerful Radeon RX 9070 graphics, and liquid CPU cooling. Comes with Windows 11 and Microsoft Office installed.
- Operating System: windows
- Tutorial Link used: I NEED ONE
tepid charm
can any1 send me a voice changer link
nimble girder
can anyone help me test my new ai chatbot site?
viral mason
- Goal making a voice
- Specific Issue: idk how gng
- Full GPU Name: AMD Ryzen 9...
Yo I saw your friend request, I can help with this in the morning when I wake up bro
I'll be able to help you in the morning when I wake up, DM me in about maybe 10 hours or so
It's 1 am for me rn
forest rose
Hi everyone, i have been working on a new AI project and i think it's finished, but I would love if someone could test it and share their thoughts, thanks!
low shard
can any1 send me a voice changer link
!help-template
patent trellisBOT
!help-template
low shard
- Goal making a voice
- Specific Issue: idk how gng
- Full GPU Name: AMD Ryzen 9...
Read up the applio ai hub docs guide
-rvc
low shard
Do i ask here for help?
!help-template
patent trellisBOT
!help-template
glacial bear
- Goal: Real time voice changer for Roleplay
- Specific Issue: I was able to run the okada fork which belonged to AI HUB before but i have started to face issues with the voice changer it is perfectly changing on server option but gets squeaky or noisy when i try to run it on the client option.
- System Specs: Running of a Apple M4 air
- Tutorial Link used:https://docs.aihub.gg
proven hill
- Goal: Real time voice changer for Roleplay
- Specific Issue: I was able to ru...
wich voice changer are you using?
west temple
im getting 190 perf, using 196ms only with 2.7 sec
anyone help?
brazen holly
- Goal (e.g., TTS, AI Covers, Roleplay): Create a voice model
- Specific Issue: I don't know how to start
- Full GPU Name: RX 6700 XT
- Operating System: Windows 11
- Tutorial Link used: None
lyric fossil
- Goal (e.g., TTS, AI Covers, Roleplay): AI RVC Roleplay
- Specific Issue: already selected input device, but still asking for one
- Full GPU Name: Tesla x2
- Operating System: Kaggle
- Tutorial Link used: https://docs.aihub.gg/realtime-voice-changer/local/tg-develops-w-okada-fork/
lyric fossil
- Goal (e.g., TTS, AI Covers, Roleplay): AI RVC Roleplay
- Specific Issue: alrea...
maybe anyone know which setting i've missed, why does it still ask for my audio input device when i already have one?
proven hill
- Goal (e.g., TTS, AI Covers, Roleplay): Create a voice model
- Specific Issue: ...
use the docs
spiral saffron
who can help me get visuals for my tiktok
<@&1159293204038955078>
viral mason
- Goal (e.g., TTS, AI Covers, Roleplay): Create a voice model
- Specific Issue: ...
I can help with that, I've got tutorial videos on how to clean audio for a dataset and how to use Applio on Kaggle to train a model
spiral saffron
<a:WEEEEEEEEE:1450152809399058576>
viral mason
what are you trying to use the voice changer for? btw that video is very outdated
all yt tutorials are old
void flume
what are you trying to use the voice changer for? btw that video is very outdate...
Note: they're using one of wido's videos apparently.
tags of concern: #trolling #egirl #roblox
also the thumbnail is concerning, but not as much as their latest
viral mason
how so? that video says "egirl ai voice changer"
that isn't allowed if you're using it for that reason
void flume
<@&1159293140440723499> (this may need investigating / judgement)
The video's transscript shows the following:
that's catfishing and whatnot
viral mason
no accusations, just saying if that's the reason then no help
people often come from those werid tutorials and want to do bad things, so it's common to ask
void flume
It's just concerning. I'm not accusing you. Just saying its sus.
viral mason
by using a normal model maybe? like Goku or something that interests you
would be funnier
Imagine u call ur friend and u answer as Spongebob
one sec, I'll get the links
here ya go, first link is the voice changer, and second link is a virtual audio cable to connect it to games or discord
u need both
for the virtual cable run setup64 and then install driver
and after for the voice changer run mmvcserversio
once u install it I can help with that
not currently
the program isn't difficult to use tho
viral mason
for the virtual cable run setup64 and then install driver
u did these two things right?
yep that file
👍
did u extract the voice changer btw
both need to be extracted before running anything
lol
it's downloading everything needed for it to work, then it should open up in browser
karmic trellis
Hi, how much does my microphone quality itself matter? I currently have a one year old headset and I use the microphone of it (its not trash, but not good either for sure), but i will soon get a dynamic mic. Will it improve the quality of the voice models? I sound a bit robotic sometimes, and it makes weird noises sometimes, no matter what settings i use in vonovox.
viral mason
it's stuck?
try clicking the command window and hitting enter a few times, idk if that helps but interacting with it sometimes does stuff for me
cool it loaded
u use windows btw right?
analog obsidian
Hi, how much does my microphone quality itself matter? I currently have a one ye...
the input (your voice) gets resampled to 16khz, that is cellphone quality, then the model uses the phone quality audio to extract the phonemes and do the voice conversion
having a better microphone helps the model stabilizing its outputs, leading to better pronunciation, and in some cases better quality too
is every model you tried robotic? if other models sound better it means the model you're currently using was trained using a bad dataset and perhaps it's overtrained
viral mason
almost
mainly audio settings
run setup64.exe
not the one with the a
then install driver
make sure in ur audio settings over to the left to set your regular mic back to default
you should be
so in the voice changer make your input your regular mic, then output your new virtual microphoen (line 1)
the what?
set your gpu to your AMD card, and fix the input and output if u need to
it's easier to check using discord before going to other ppl
do u have the settings right on other apps?
karmic trellis
the input (your voice) gets resampled to 16khz, that is cellphone quality, then ...
Every model is robotic for me, and i have tried a lot. I also used vonovox audio player to test out my voice models with other peoples voices from random english YouTube videos i recorded. Then the random "weird noises" and the "robotic feeling" was gone. I dont think the problem is the voice model itself, because in that case the voice models should be bad with other peoples voice as well, right?
analog obsidian
Every model is robotic for me, and i have tried a lot. I also used vonovox audio...
i see, i think vonovox has a noise suppressor vst, but if you already tried that then yea it could be that your current headset microphone is not clean enough for the model to understand what are you saying
new mic should in theory fix it
median monolith
How do I inference my audio (Applio Google Colab method) without the automatic "Instrumental" separation? just "cover" the audio as it is
stoic prairie
I have a 4080 Super. What real time voice changer should i use?
keen dirge
- Goal (e.g., TTS, AI Covers, Roleplay): AI cover
- Specific Issue: replay isnt outputting converted audio
- Full GPU Name: nvidia 4070 ti
- Operating System: windows 11
- Tutorial Link used: none
it worked just fine like a month ago and now all of a sudden its not working anymore
low shard
- Goal (e.g., TTS, AI Covers, Roleplay): AI cover
- Specific Issue: replay isnt ...
share the downloaded program you’re using
patent trellisBOT
!help-template
keen dirge
share the downloaded program you’re using
Replay
low shard
I have a 4080 Super. What real time voice changer should i use?
!help-template
patent trellisBOT
!help-template
low shard
Replay
Replay is unsupported, it’s abandoned since weights shutdown forever, use Applio instead
-rvc
plucky crown
If you got a supported gpu then the gpu
Is there any voice models i can download for applio? Is there a website for that or here?
low shard
Is there any voice models i can download for applio? Is there a website for that...
#1175430844685484042 , its better you elaboratr
deep raft
hi i download ollama and its run good but i want to know how i could learn my model about specific data on windows 10
plucky crown
<#1175430844685484042> , its better you elaboratr
It says select voice model but i dont have any ready made , is there any best trained voice model here for appolio so i can test how this applio works
plucky crown
Hey is there a number of audio and audio length limit in applio voice training?
jagged blade
what the recommended amount of training
low shard
It says select voice model but i dont have any ready made , is there any best tr...
you need to use the hep template
hardy belfry
Hi everyone, I’m trying to create an English version of a Japanese song while keeping the original instrumental, melody, tempo, and structure.
What is the best possible way to it? Thanks
proven hill
Hi everyone, I’m trying to create an English version of a Japanese song while ke...
for free?
i dont think theres free programs
hallow thistle
i'm using kaggle
Applio RVC notebook?
young halo
Applio RVC notebook?
yep
fiery plinth
Does anyone know what sort of prompt/site/software people use to make those AI brainrot reaction images that are all over IG Reels? Those ones that are all like "when I step on the mango mustard golden dandelion 67 CITY BOII" or whatever
long rose
Hey is there anyone with just 5 minutes to get into a call. my head is so messy right now cuz im getting so frustrated, if someone would be willing to help me id really appreciate it
sullen mantle
why does the sliced audios sound bad quality
When I listen to them
but my actual dataset sounds higher quality
anyone know
hallow thistle
Does anyone know what sort of prompt/site/software people use to make those AI b...
ChatGPT's image model can be used to make brainrot images as far I know, but Google Gemini's Nano Banana and some other image generators are the honorable mention as well.
fiery plinth
ChatGPT's image model can be used to make brainrot images as far I know, but Goo...
I posted some reference pics in the thread I started. Will start with taking some of the original/base pics to plug in as references and then prompt for things such as lightning and laser eyes lol
torpid igloo
i need someone with chatgpt pro
hallow thistle
<a:PaiNod:889680917348622377>
I posted some reference pics in the thread I star...
Good luck. If you're interested to show your own generation, you can send to #🏙│ai-images, but you'll get an image perm once your username turns blue here. 
hallow thistle
Hey is there anyone with just 5 minutes to get into a call. my head is so messy ...
What is it about?
long rose
What is it about?
So im having so much trouble with cognitive overload, i cannot think straight, mainly because my interests are so many, and i jump all over the place. Ive tried physical planners an more, but i just cant seem to map it out clearly. Therefore ive started working with AI, such as adaptive.ai and perplexity computer, but there is just so much going wrong and its so frustrating. My goal is to make a basic AI agent / assistant who can remind me, relief some of the stress and give me clear structure
viral mason
that isn't allowed <@&1159293140440723499>
sullen mantle
that isn't allowed <@&1159293140440723499>
yo
void flume
that isn't allowed <@&1159293140440723499>
Thanks
viral mason
how to fix high perf
if u have a 4060 u should be using Vonovox for realtime
also if u are using it in games turn down the graphics of the game to see if that improves performance
distant sable
why is it glitching and sounds robotic
viral mason
why is it glitching and sounds robotic
not sure, what GPU do u have? (Nvidia or AMD) and what are u using the voice changer for
viral mason
3070 nvidia
ok cool, u should use Vonovox, you probably downloaded something off a very old youtube tutorial, what kind of voices are u using btw, Anime characters or maybe Spongebob ect
distant sable
ok cool, u should use Vonovox, you probably downloaded something off a very old ...
I use okada and yeah lets just say spongbob
viral mason
wdym let's just say?
distant sable
wdym let's just say?
like I use it
viral mason
you're not using it for anything weird right?
distant sable
you're not using it for anything weird right?
wdym?
viral mason
I use okada and yeah lets just say spongbob
idk u made it sound sus lol
distant sable
nah just for trolls
viral mason
eh?
distant sable
alr what do I use
viral mason
wdym troll
distant sable
what voicechanger should I use
viral mason
this is the newest and best realtime voice changer for Nvidia
second link is a virtual audio cable like VB cable to connect it to discord or games
if u have one already u don't need it then
distant sable
Vonovox! https://huggingface.co/dr87/vonovox/resolve/c8034f5f6d50648a8109bb4f847...
how good is it?
viral mason
it's the best currently. Quality is much better, pretty much no robot sounds unless you're using an old model that wasn't trained good, and it's very fast
distant sable
it's the best currently. Quality is much better, pretty much no robot sounds unl...
how do I open it
viral mason
run the file called Start, make sure to extract it
foggy brook
does anyone know how to get a better ollama UI than cmd?
simple ore
does anyone know how to get a better ollama UI than cmd?
LM Studio?
viral mason
what kind of trolling?
I mean what kinda models are u using
like anime, Spongebob, something cool like that
oh
it's not allowed to do that here <@&1159293140440723499>
should have read the rules
valid axle
Okay, so I noticed that all, not most, ALL online AIs are greedy paid services with requiring paid monthky plans to even use them properly, and even if local AIs exist their quality aint as good AND worst of all it requires 64GB+ RAM PCs which obviously not a lot of people can afford.
As someone who wants to use AI as a tool without paying yet has a modest gaming PC what is the best way to use AI for game dev and animation(bonus points if I can train AIs as well as use references)
viral mason
What do u guys use to train vonovox and applio is it mp3 or wav
what?
viral mason
Vonovox is only a voice changer to use the voice models, Applio is what is used to train them
plucky crown
Vonovox is only a voice changer to use the voice models, Applio is what is used ...
In mp3 i was able to download a 30minute video in 64mb
But in youtube to wav online converter the result is 285mb
Which should i pick
viral mason
mp3 aren't suggested, it's best to use Wav and download with yt dlp
file size isn't what matters it's quality
plucky crown
file size isn't what matters it's quality
I noticed the video's audio has like background noise, whats the best do u guys use to clean the voice or having background noise is recommended?
viral mason
I personally use this site https://mvsep.com/en/home
I made a cleaning tutorial on how I usually go about my models
plucky crown
I made a cleaning tutorial on how I usually go about my models
Whats your youtube channel linkk
viral mason
this is the tutorial I don't use yt for much
plucky crown
I made a cleaning tutorial on how I usually go about my models
That is the best one right? Can i use that offline like local download
plucky crown
no it's a website, online only
For training what epoch is the best quality
viral mason
there isn't one, just train it until it sounds good
plucky crown
If i export the trained one can i retrain it again with more audio if the result export wasnt good
viral mason
wdym?
plucky crown
I personally use this site https://mvsep.com/en/home
Is there an alternative that is free or has free credits/trial
viral mason
Is there an alternative that is free or has free credits/trial
uvr is free, also mvsep is free too
plucky crown
uvr is free, also mvsep is free too
But for mvsep the export to wav needs credit
Oh wait
Its only the wav32bit
But now my wav file it says file is too large
viral mason
But for mvsep the export to wav needs credit
it doesn't tho
convert the audio to flac
plucky crown
Is uvr as good as mvsep
Also for the voice model from applio
If i export the first trained voice model can i retrain it again with more audio if the result export wasnt good
plucky crown
I personally use this site https://mvsep.com/en/home
Also what do u use to download youtube links to wav?
viral mason
Also what do u use to download youtube links to wav?
I use Yt Dlp
plucky crown
I use Yt Dlp
What command do u use via cmd to download the wav file of a youtube link?
viral mason
Is uvr as good as mvsep
yt-dlp.exe -x --audio-format wav yt link
it requires the app tho so you need to download it
livid lodge
Goal: Roleplay, Specific Issue: Not seeing virtual devices, GPU: Nvidia 4090, OS: EndeavorOS (Arch Linux), Tutorial link used: tg-develop rvc setup. More detail: I'm trying to get this to work in arch linux, and while I feel I'm almost there, there's just no sound that ever comes out of here and I don't have any virtual audio devices showing up in discord.. I've tried everything I know how, but I'm surely doing something wrong here. I followed all the instructions on the site for the tg-develop rvc setup (cuda) but am just stuck now because i think its just not hearing my mic, and it's not seeing my virtual cable despite pipewire being installed. Any ideas?
void flume
I'm not too familiar with linux myself, so this may not be very helpful
Can you see if the microphone is working outside of the voice changer?
Does the browser have permissions to use the microphone and speaker?
Of course, is it possible to use the virtual cable you created with pipe wire at all, as in; did you test it outside of pipewire and the voice changer?
Basically, just try testing things like and narrow down under what exact circumstances things stop working, because then it will be easier to fix.
void flume
Goal: Roleplay, Specific Issue: Not seeing virtual devices, GPU: Nvidia 4090, OS...
I should add that, one thing I don't understand about the guide is that, for linux it makes people download portaudio, however you can already 'wire' audio (send it from one application to another) using pulseaudio which is in most linux distros. Also this utility will let you work with that more easily https://github.com/theRealCarneiro/pulsemeeter
ocean maple
where can i download the lastest vonovox softare
viral mason
where can i download the lastest vonovox softare
It's best to get it from the official discord server but this is the most recent beta, much better than the current full release
https://huggingface.co/dr87/vonovox/resolve/c8034f5f6d50648a8109bb4f847182362e2b779b/Vonovox_beta_17_11.zip
livid lodge
I'm not too familiar with linux myself, so this may not be very helpful
Can you ...
Hi Rumi, sorry Discord did not ping me. I got this working actually, just now. I am not sure why but I had to completely reset my config. I destroyed all the mics/etc and started with a clean slate and it worked. Portaudio, in my case, wasn't necessary at all since my distro already had the capability here. I did have to create a virtual mic for discord though. I used Helvum to view things better which also helped. So sorry for the earlier message. Perhaps I should have tried to just start from square one again. XD
ocean maple
It's best to get it from the official discord server but this is the most recent...
hey can i ask u a question aobut vonovox
how can i use like the models i find here
wait nvm i figured it out
i think
plucky crown
Helpppp what model do you guys use in UVR5 to get the cleanest vocal with no background noise?? Latest
orchid horizon
I need music/audio extend tool
I using RTX 2080 16gb
and I searching extender with no credit no limit and uneditable lyrics (you cant extend with your own lyrics) at the same time my friend has it and him dont give it
void flume
I need music/audio extend tool
I using RTX 2080 16gb
and I searching extender w...
Yeah, idk any. I did a quick search, and all of them have some sort of limit. Example: https://github.com/ace-step/ACE-Step-1.5 (limit: 10 minutes)
I recommend waiting for when more people are online.
plucky crown
How would you know if it gets worse
So does that mean you can retrain a smaller epoch to make it bigger epoch?
fringe heron
What sampling rate do i pick then
32khz
real tinsel
-issue: My issue is my terminal is not running it just shows text, I have a RTX 5070 TI 32 GB Windows 11 and used this tutorial https://www.youtube.com/watch?v=agDg2kDEwX4
hallow thistle
-issue: My issue is my terminal is not running it just shows text, I have a RTX ...
What are you trying to use the voice changer for?
real tinsel
To change My voice for my friends and play around with it
hallow thistle
To change My voice for my friends and play around with it
You're following an outdated tutorial video that tells you how to use the older voice changer with "E-girl voice model", and just letting you know using E-girl model is against the server rule if you're wondering.
real tinsel
You're following an outdated tutorial video that tells you how to use the older ...
I cant even use it in different servers ?
Im just here to set it up and leave
hallow thistle
Well, there are options. Either try more recent voice changer and use another funny voice model. 
real tinsel
Well, there are options. Either try more recent voice changer and use another fu...
is there any other videos that walk through it i cant find any new ones
hallow thistle
Better not to follow any tutorial from YouTube about voice changer. There's Vonovox.
real tinsel
how should I set that up
plucky crown
32khz
Helppp i have i7 13700 rtx 3090 what chunk length also do i check noise filter and noise reduction (i already processed the 41 10-30 minute audio files in uvr5) also normalization mode which do i pick none,pre or post? Also overlap length? I want the best best voice quality possible
@hallow thistle
plucky crown
Helppp i have i7 13700 rtx 3090 what chunk length also do i check noise filter a...
For batch size i should go for 24gb right cuz i have 24gb vram? Also save every epoch at what? Total epoch set to what? I have 41 wav audio files also do i check the cache dataset to gpu? How about overtraining? What do i check in advanced settings
fringe heron
Can help in an hour sorry
fringe heron
Guys what batch size plsss
16 is enough
plucky crown
16 is enough
In advance setting what else do i check , also i set epoch to 1000
fringe heron
In advance setting what else do i check , also i set epoch to 1000
Higher epoch doest mean greater quality
plucky crown
Do i check fresh training? Cache Dataset to gpu? How about overtraining detector?
fringe heron
Do i check fresh training? Cache Dataset to gpu? How about overtraining detector...
No, just overtrainer
plucky crown
Higher epoch doest mean greater quality
What do i pick then they said they play around the epoch numbers to see whats best so if i set to 1000 and it saves every 10 does that mean i can check lower epoch results or nah retrain it again with a new epoch number
fringe heron
What do i pick then they said they play around the epoch numbers to see whats be...
Each 10 it saves a model and you can hear
plucky crown
So what epoch number should i set it as
fringe heron
1000 and overtrainer ypu wont even finish them
plucky crown
Helppp i have i7 13700 rtx 3090 what chunk length also do i check noise filter a...
This is what i have 41 10-30minute wav audio mostly in 25-30 minute each though
plucky crown
What do i do then what epoch do i set it
fringe heron
What do i do then what epoch do i set it
What
hallow thistle
Set total epoch at least 500.
plucky crown
1000 and overtrainer ypu wont even finish them
You said if i pick 1000 i wont finish it , does that mean the training would take forever
fringe heron
No
plucky crown
Like weeks sjgdians
fringe heron
It ends earlier
plucky crown
You might not finish 500
What does might not finish mean
hallow thistle
You might not finish 500
I don't understand this one.
fringe heron
It ends earlier
If you Set 500 and overtrainer to 50 if after 50 epochs ther is no improvements it ends
fringe heron
I don't understand this one.
Sorry for any misunderstranding
Its hard to text for me rn
plucky crown
If you Set 500 and overtrainer to 50 if after 50 epochs ther is no improvements...
alright thank you so much!!!
How about index algorithm should i set it to auto?
fringe heron
How about index algorithm should i set it to auto?
You prob wont use it
Its up to you
plucky crown
Do i generate index first or nah start training
fringe heron
Helppp i have i7 13700 rtx 3090 what chunk length also do i check noise filter a...
Also prob you already split but splitting in the software you are using doest even matter
plucky crown
Do i generate index first or nah start training
How about thiss
fringe heron
How about thiss
Start Training
Index, if you have large datasets is even more useless
plucky crown
Does this need internet?? Why did it took a second
fringe heron
No it doesnt
fringe heron
Does this need internet?? Why did it took a second
For what?
plucky crown
fringe heron
What took a sec
plucky crown
It says this immediately once i clicked start training
It says trained successfully
fringe heron
You have a cmd opened?
plucky crown
Yes
plucky crown
plucky crown
The tutorial said i should zip it
fringe heron
The tutorial said i should zip it
You should not
plucky crown
Alright wait upp
fringe heron
Sure thing
fringe heron
Alright wait upp
Also since you are aiming for max quality there may be more things to cover
plucky crown
Also since you are aiming for max quality there may be more things to cover
Im preprocessing audio its saying this what does it mean
Its saying errors
plucky crown
It now says this
fringe heron
Yea but answer
fringe heron
Okay
plucky crown
I thought it was 41
plucky crown
How do i do that
plucky crown
It now says preprocess completed in 55.26 seconds on 08:19:11 seconds on audio or nah still split it
Yeah i can use python code to split the audios, you mean it needs the length to be smaller
fringe heron
If there something was processed is not complete
fringe heron
Yeah i can use python code to split the audios, you mean it needs the length to ...
Yea
plucky crown
What length should i cut the audios then so there will no longer be errors
fringe heron
What length should i cut the audios then so there will no longer be errors
Ram you have?
plucky crown
31.8 gb ram
plucky crown
fringe heron
Yea i need to guess by graph like
It doesnt show how much used in numbers
Can you pic better
proven hill
why dont you take screenshots?
fringe heron
Thanks
fringe heron
why dont you take screenshots?
He sends stuff, at this point i ask what i need
fringe heron
why dont you take screenshots?
Is a problem?
proven hill
Is a problem?
no but i mean
plucky crown
Well you go with clips long 5 mins each
Alright 5 minutes seperated in all audios then, with my 3090 gpu i7 13700 how long do u think it might take to train 42 wav audios ranging 10-30minutes but mostly 25-30min seperate 5 minutes each? Cuz its night here right now i hope my pc wont blow up while i leave it on running while sleeping
It 7:23pm right now here
Theres like video online about gpus like catching fire on its own so aaa would many wav files can cause it or nahhh
fringe heron
Alright 5 minutes seperated in all audios then, with my 3090 gpu i7 13700 how lo...
It wont blow
fringe heron
Theres like video online about gpus like catching fire on its own so aaa would m...
Dw
fringe heron
Alright 5 minutes seperated in all audios then, with my 3090 gpu i7 13700 how lo...
A bit
It will take a bit
fringe heron
Alright 5 minutes seperated in all audios then, with my 3090 gpu i7 13700 how lo...
You will use the model on your voice?
plucky crown
You will use the model on your voice?
Yes realtime voice too, im gonna do all the voices on creating music im very broke i gave all my savings to the pc
fringe heron
Man goodluck
fringe heron
Yes realtime voice too, im gonna do all the voices on creating music im very bro...
It might not sound as good as you hope
plucky crown
It might not sound as good as you hope
Really what does that meann
fringe heron
Really what does that meann
Your voice means a lot for quality
Lets say your voice is far away from target it might sound off
plucky crown
Yes, so is applio trained voice and then put the trained file to vonovox will the sound be accurate
fringe heron
Yes, so is applio trained voice and then put the trained file to vonovox will th...
Maybe
To my exp
plucky crown
Can u send me a sample , now im nervous
plucky crown
Alright thank you
Also example if i trained the file can i retrain it again with new audio files if i am not satisfied with the sound it creates
fringe heron
Also example if i trained the file can i retrain it again with new audio files i...
Yea
I can provide way better help later tho
plucky crown
It might not sound as good as you hope
Can i still get my hopes up by spamming more audio files so it can level up to adjust if the singing involved screaming and many more quick tone changes, cuz voice quality matters alot to the listener especially in music, as long as the voice sounds different from mine so it sounds like different people singing in the music
fringe heron
Can i still get my hopes up by spamming more audio files so it can level up to a...
Meh
plucky crown
What does that mean
plucky crown
Which is more convincing those ai tts or the vonovox one
fringe heron
Which is more convincing those ai tts or the vonovox one
Vonovox
But the model itself does the lifting
plucky crown
Really but ai tts are like so advanced now only small audio files and now theyre convincing
Is vonovox really better
Is vonovox also what those vtubers used
fringe heron
Is vonovox also what those vtubers used
Idk
fringe heron
Is vonovox really better
Once again, model does the heavy lifting
If that doesnt work well on your voice the model is the bottleneck
Like for me, its a pain, i am not even a native english speaker
My speaking style makes model sound a bit off, almost anything
Any target pretty much
I have tons of samples
They sound perceptually not right
plucky crown
I can provide way better help later tho
Chat with you soon then imma make a python code for cutting the audios to 5 minutes
fringe heron
Chat with you soon then imma make a python code for cutting the audios to 5 minu...
Okay
fringe heron
Chat with you soon then imma make a python code for cutting the audios to 5 minu...
Ping me when you will thought
plucky crown
Ping me when you will thought
Is this normal
fringe heron
Is this normal
Yes
plucky crown
This is for just 1 5 minute wav? I produced 180 5 minute (and others less then 5minute) videos
fringe heron
This is for just 1 5 minute wav? I produced 180 5 minute (and others less then 5...
You mean audio?
What video
plucky crown
The wav file , yes
fringe heron
This is for just 1 5 minute wav? I produced 180 5 minute (and others less then 5...
No thats not for 1
1 epoch = ai saw full dataset
plucky crown
Whats epoch2? Is this like the same 5 minute wav and it will count like up to 1000 (i set the epoch to 1000 out of curiousity)
fringe heron
Hold on there is a misunderstanding
fringe heron
Whats epoch2? Is this like the same 5 minute wav and it will count like up to 10...
5 min video what
Like is the dataset made out of all the files each long 5 mins?
I guess so no?
plucky crown
Yes the 42 wav files ranging 10-30 minutes turned to 5 minute (and the left over less than 5 minute) cuts
fringe heron
Okay
Yea it eill count to 1000 epochs but since you have the overtrainer on it mostlikely wont count up to 1000
You set it to 50
plucky crown
Yes i set it to 50
fringe heron
So after 50 epochs of 0 improvements it will end regardless of the 1000
Like, 430 epochs, you had 0 improvements over 50, it will end at 480
plucky crown
Is this normal
Do you also train locally too? Is this gpu temp normal? Hopefully it doesnt go higher than that since i saw videos of like people's gpu smoking all of a sudden
fringe heron
Do you also train locally too? Is this gpu temp normal? Hopefully it doesnt go h...
Ye its fine
plucky crown
I just saw it go to 80°C
fringe heron
90 would be odd but 80 is okay
Thought wait
May you tell me vram usage please
Just write it
plucky crown
proven hill
why dont you take screenshots tho i dont get it
plucky crown
How to see vram usage
fringe heron
why dont you take screenshots tho i dont get it
Same
fringe heron
Whats epoch2? Is this like the same 5 minute wav and it will count like up to 10...
Also if you want you can already listen to a sample, it saved 1 model
plucky crown
So meaning now its in epoch 3 that means one epoch is a whole 180 wav files? And it has to go to like epoch 200 or 500 or wherever it feels the training is enough?
fringe heron
So meaning now its in epoch 3 that means one epoch is a whole 180 wav files? And...
Yes
Also you asked earlier if i train locally too and the answer is yes
plucky crown
It takes like 6minutes to 7 to finish one epoch hopefully when i sleep this pc doesnt blow up
fringe heron
It takes like 6minutes to 7 to finish one epoch hopefully when i sleep this pc d...
No it wont
You can sleep safe
plucky crown
No it wont
Whats the max temp did you gett
fringe heron
Old gpu, 80 (3060 12 gb)
plucky crown
No it wont
You say that but your pfp says something else lol
fringe heron
New gpu, 70 (5070 ti 16 gb)
Never really reaches that temp tho
Max usually is 60
Important: only becouse yours is higher doesnt mean its bad or harmful
plucky crown
I saw people in other forums like has latest 5060 series gpu that has 8gb and they said they dont think they can use these applio and stuff
fringe heron
I saw people in other forums like has latest 5060 series gpu that has 8gb and th...
Idk about them
Sounds wierd
Anyway the dataset is a single person right?
plucky crown
Anyway the dataset is a single person right?
Yes
plucky crown
Yes
fringe heron
Cool
If you do music with it please pay attention, i dont know where you get it from, just telling you
plucky crown
Also i heard about this seedvc vs vonovox one what do u think , im seeing this in like youtube comments
fringe heron
And you publish ofc
fringe heron
Also i heard about this seedvc vs vonovox one what do u think , im seeing this i...
Trust me, seed vc is not better
I tried it and its, quite bad if you want max quality
(If i remember right, i tried many things)
plucky crown
Trust me, seed vc is not better
Whats your top 3?
fringe heron
Whats your top 3?
Rvc (v2), thats the only worth mentioning tbh
And I dont remember the overall quality of other as its been quite a bit, unless they updated stuff you should use this
What you are already using
plucky crown
Rvc (v2), thats the only worth mentioning tbh
Is rvc v2 the same thing as vonovox they do realtime voice and voice file to changed voice?
fringe heron
Is rvc v2 the same thing as vonovox they do realtime voice and voice file to cha...
See vonovox as an engine that can run your models, yes it supports the v2 variant
fringe heron
So meaning now its in epoch 3 that means one epoch is a whole 180 wav files? And...
Good luck getting it to sound good
Also with 17 hours you might take a whole day
Maybe 20 hours
plucky crown
Rvc (v2), thats the only worth mentioning tbh
Whats the link to this, the only ones they kept mentioning to me is applio and vonovox
fringe heron
Whats the link to this, the only ones they kept mentioning to me is applio and v...
In short you are already using the best
fringe heron
So meaning now its in epoch 3 that means one epoch is a whole 180 wav files? And...
Also about training you might want to use a tool to look how training is doing
And its called tensorboard
plucky crown
And its called tensorboard
Is it the one in the applio file the run tensorboard one
fringe heron
Is it the one in the applio file the run tensorboard one
Yea
Doesnt work for me thought
So i do it manually
If for you it doesnt work i raccomand to just install tensorboard from pip and simply then running in the cmd "tensorboard --logdir=[path to the training folder]"
fringe heron
If i click it it wont
It wont what
fringe heron
To me looks like is running fine
fringe heron
Can you go over scalars and look there
Also no need for all images
You could if really needed just 1 of all the scalars
fringe heron
May you write the numbers please and not all images
plucky crown
Also already NaNs
What does that mean is it bad
fringe heron
In norm_g
You can see in the graph 2 arrows going up
At the beginning
If you hover eith the mouse it should say NaN, you should be good
Also i raccomand to smooth the graphs (in scalars all the way to the left)
Set it to 0,987 or 0,999
plucky crown
What do i pick it says 0.6 also if we change scalar is it affecting the training
plucky crown
plucky crown
obsidian thunder
Hey, it's been a long time since I used AI voice models, and I've forgotten most of it.
What's the best way to export voice models to an MP3 file?
fringe heron
Hey, it's been a long time since I used AI voice models, and I've forgotten most...
Export a model to an mp3?
obsidian thunder
translator was tweaking xd
obsidian thunder
Export a model to an mp3?
i meant put an voice model on an mp3 like voice lines from an song or smth and export it to mp3
fringe heron
i meant put an voice model on an mp3 like voice lines from an song or smth and e...
Converting the vocals of a song with a model then puntting it on top of the track?
obsidian thunder
yeah but i dont know anything anymore how to convert the vocals with a model like programs websites like that xd
fringe heron
yeah but i dont know anything anymore how to convert the vocals with a model lik...
I would like to help but i dont have much time sorry :(
hardy yew
i meant put an voice model on an mp3 like voice lines from an song or smth and e...
So, converting sound files to sound files with RVC? Use Applio.
limber sundial
Yes I read
desert notch
hey guys anybody know how to get a indian female rvc model or voice
low mason
why whenever i open the MMVCServerSIO it open a website
how do i open the application/pop up i dont want to use the website
viral mason
why whenever i open the MMVCServerSIO it open a website
Most likely you have either wokada deiteris or Wokada tg fork
They only open on browsers but they work the same
proven hill
Most likely you have either wokada deiteris or Wokada tg fork
yea and that cant be changed
viral mason
Ye
If they have AMD they're stuck with it on browser since no Vonovox
I don't get the problem tho besides I guess the browser eating up gpu or cpu
Especially if it's google
low mason
Most likely you have either wokada deiteris or Wokada tg fork
do u have the download link off the other version that doesnt use website
viral mason
Do you have an Nvidia gpu?
low mason
yea
viral mason
Is it a 1660, 3070, anything like that?
low mason
4060 RTX
viral mason
Ooh ok, Vonovox runs on an app and is much better than those two I mentioned
I'll get the links rq
low mason
is it real time voice changer aswell vonovox?
viral mason
is it real time voice changer aswell vonovox?
it is yes
u can use the same models u used in the other one
second link is a virtual audio cable for connecting it to games or discord, if u have one already u don't need it
thick canyon
could someone help me with wokada?
viral mason
could someone help me with wokada?
what gpu do u have (Nvidia or AMD) and what are u using the voice changer for?
thick canyon
what gpu do u have (Nvidia or AMD) and what are u using the voice changer for?
3080, vr chat
i clicked on start but no one is hearing me
viral mason
what kinda models do u wanna use? like Venom, Anime characters, Spongebob ect
thick canyon
girl voices
viral mason
why?
thick canyon
wanna test these
viral mason
why girl voices in particular tho, any reason?
thick canyon
if its better than dubbing ai or not
i have a voice disorder and doesnt like my regular voice
viral mason
Vonovox does the same thing but it's much better optimized, better quality and such
thick canyon
Okay
but it still doesnt fix my issue
nothing comes out when i select virtual audio cable
viral mason
are you able to send a screenshot?
thick canyon
https://huggingface.co/dr87/vonovox/resolve/c8034f5f6d50648a8109bb4f847182362e2b...
i need download both?
viral mason
the second link is a virtual audio cable so if u have one already u don't need it
thick canyon
i have it already
viral mason
alrighty
thick canyon
downloaded it
viral mason
your mic should be your default mic, is that is
thick canyon
your mic should be your default mic, is that is
yes it is
some settings i need to do?
and adding models
pallid patio
Good night! Where can I find the link for the collabs?
viral mason
Good night! Where can I find the link for the collabs?
-colab
viral mason
here ya go good sir
thick canyon
where i can find good models
thick canyon
ok ty
viral mason
you're welcome :D
thick canyon
you're welcome :D
the Real E-girl (RVC 2 - 60 Epochs) voice
viral mason
:(
thick canyon
i saw it got updated
viral mason
use something normal instead of an egirl voice
thick canyon
use something normal instead of an egirl voice
i try it for first
just need an start base
viral mason
there are normal female voices on female characters
thick canyon
you know how to use the newer version?
thick canyon
there is old and new
viral mason
old and new what
thick canyon
viral mason
I don't use egirl models. so idk
thick canyon
you're welcome, please use it responsibly
will do! just unsure if i should use the effects like noise gate
is gpu delay 60 ms good?
viral mason
I've got my vonovox like this, I at times use noise gate but it's very sensitive depending on the model
other effects I use in FL studio
which would require a complex setup
viral mason
is gpu delay 60 ms good?
I think so
if u turn the graphics of vrc down some it could increase preformance
I play the game as well and could test to see if u got it working btw, my user is the same here as on vr
viral mason
What does index do?
basically the index makes the model sound more like itself
thick canyon
basically the index makes the model sound more like itself
aah thank you
covert zenith
hi all , i am hopping to get some advice on how to connect my main computer to my RVC computer :
both are connected to the same network
RVC computer specs is 8700K with rtx 2080 super
i am using deiteris voice-changer
viral mason
hi all , i am hopping to get some advice on how to connect my main computer to m...
u should use wokada tg fork instead unless you've gotten bad results from he lower gpu on that pc
thick canyon
Quality from vr chat to discord makes so much difference haha
covert zenith
u should use wokada tg fork instead unless you've gotten bad results from he low...
i see , thank you for the reply . i was not clear when asking . i am trying to connect my main pc to my second pc which will be running the rvc but do not know how to
viral mason
i see , thank you for the reply . i was not clear when asking . i am trying to c...
ah, I'm not sure about that in particular, try askin one of the helpers by pinging the @ helpers role
covert zenith
ah, I'm not sure about that in particular, try askin one of the helpers by pingi...
ah i see , thank you very much
hi @tame oracle , i am hopping to get some advice on how to connect my main computer to my RVC computer :
both are connected to the same network
RVC computer specs is 8700K with rtx 2080 super
i am using deiteris voice-changer
are you free and able to advice me ?
brazen holly
hey, what's the website y'all use to use voice changer?
i know the voice changer in real time but not like premade
brazen holly
i know the voice changer in real time but not like premade
or even voice changer in real time, what do y'all use please?
sacred mica
hey I was wondering if I want to do TTS, should I go with Fish s2? Currently running a gguf version of it on a 2080 ti 11gb vRam; I don't mind spending money on runpod or smt to train models but I'm trying to accurately create TTS voices that have human emphasizes voice acting like almost
Its js getting the prompt right for it seems impossible is that the best way to goa bout doing it
also is there a reverse speech to text? that can tell me what tags a certain audio would fall for
so i can learn wat i need to do
tyty
plucky crown
Does weights website use vonovox? I recall back then the voices there are very well made
plucky crown
Set it to 0.987
Hello an update its counting down now
It says g/total:24
If it goes to zero does that mean it got the best quality even if it reached the threshold? Its only at epoch 136
But it has been running for 16 hours now though
One epoch is like 7 minutes to 8min maxed but its at 6:30min rn per epoch
Woah 60 hours of voice? Thats more than mine! Is it very convincing and not robotic?
viral mason
Does weights website use vonovox? I recall back then the voices there are very w...
Weights has been gone for a while, it never used any of the voice changer apps
fringe heron
If it goes to zero does that mean it got the best quality even if it reached the...
You will never reach 0
fringe heron
If it goes to zero does that mean it got the best quality even if it reached the...
But yea the lower the loss is and the "best" it should sound
fringe heron
Woah 60 hours of voice? Thats more than mine! Is it very convincing and not robo...
That is an old thing and i never ended up doing
fringe heron
Does weights website use vonovox? I recall back then the voices there are very w...
You tried some of them?
fringe heron
Also looks like it may will end soon due to the overtrainer
plucky crown
You will never reach 0
Yeah
The g/total i meant like the overtrainer since if it ends does that mean it got the best sound out of the dataset?
fringe heron
Yeah
The g/total i meant like the overtrainer since if it ends does that mean ...
Probably
fringe heron
Yeah
The g/total i meant like the overtrainer since if it ends does that mean ...
You can already listen to how it sounds, just go in applio, inference tab, select your model, record real quick audio and upload it or record it from applio and listen
plucky crown
Its at g/total 16 now
fringe heron
Its at g/total 16 now
What
plucky crown
What does the numbers show now does that mean the result will be good
fringe heron
Where do you read 16
fringe heron
What does the numbers show now does that mean the result will be good
Mate maybe
fringe heron
No thats not loss
Thats the num of epoch before it ends
If there is no improvement
Use tensorboard for loss
plucky crown
fringe heron
Not the cmd
plucky crown
The scalars rightt
plucky crown
fringe heron
Dont look at norm
plucky crown
fringe heron
Smooth it
plucky crown
plucky crown
Look at avarage 50 mel and g total
Is this it
fringe heron
Yes
plucky crown
What does it say is it good
fringe heron
What does it say is it good
Mate go and listen how it sounds, trust your ears
You dont need to stop training to listen
plucky crown
If it doesnt sound well can i retrain it again with the same dataset or no its has to be different
fringe heron
If it doesnt sound well can i retrain it again with the same dataset or no its h...
It will be pretty much the same
plucky crown
You dont need to stop training to listen
But wont that break my gpu its already at 100% and 80C°
fringe heron
You can retrain and keep training in future but it wont change much
fringe heron
But wont that break my gpu its already at 100% and 80C°
For reference mine doesnt reach 80 but today is the 5th day its running ai training
Non stop
plucky crown
For reference mine doesnt reach 80 but today is the 5th day its running ai train...
Wow
How many hours of wav is it
plucky crown
But how many wav audios do you max out on each model?
fringe heron
But how many wav audios do you max out on each model?
Usually more than 10 hours atleast
15-30 or 20
plucky crown
Wow so how did the result go
plucky crown
15-30 or 20
Thats more than mine
fringe heron
Wow so how did the result go
Not that great
As i said my voice is a bit of a pain
I am working on a solution
plucky crown
Not that great
Lemme hear cmonn
plucky crown
As i said my voice is a bit of a pain
How is it a pain
fringe heron
Lemme hear cmonn
I cannot send nothing, i am on phone
plucky crown
Is the model a pain or your voice have to be close to the model's
fringe heron
Is the model a pain or your voice have to be close to the model's
The model
In short my voice is never vlose to target
So to get really high quality its very hard for me, i need to speak in some unnatural ways for me
To match the target
This is my experience atleast, prob some other will not have those issues
plucky crown
For reference mine doesnt reach 80 but today is the 5th day its running ai train...
5th day does that mean you cant use your pc for 5 days? How many gpus do u have
fringe heron
5th day does that mean you cant use your pc for 5 days? How many gpus do u have
1
Actually i have tho 3 desktops but yea, i dont use the pc a lot
plucky crown
Actually i have tho 3 desktops but yea, i dont use the pc a lot
3 desktops? What are the specs? Wow
fringe heron
3 desktops? What are the specs? Wow
Why asking specs?
fringe heron
Cuz in ai groups they have like those very expensive specs
I dont remember the cpus tho of 2 of them
They all intel all 32gb of ram, main have 5070 ti 16gb, second has 3060 12 and third a quadro k1000
plucky crown
They all intel all 32gb of ram, main have 5070 ti 16gb, second has 3060 12 and t...
Wow
fringe heron
You can use it eithout stopping
fringe heron
Also gpu got to 81C
81 degrees you mean
fringe heron
You can already listen to how it sounds, just go in applio, inference tab, selec...
Do this
winter mason
Best server settings ? not using client
west temple
if u have a 4060 u should be using Vonovox for realtime
but im kinda confused with vonovxes settings, and vonovox uses alot more vram
shy epoch
Hello guys I have a project for which I need good voice models in my language(german). It does not need to be specific ones. But no matter which platform, I just cant filter german ones out, so I joined AI HUB agein. (I thought it was still dead). Can someone help me finding actually good quality models that I can browse and try without wasting hours of time only finding different languages ? Thank you.
fringe heron
Its so bad
Some more details
void flume
Best server settings ? not using client
Please don't ignore the bot. What is your goal? What is your hardware, etc.
'best settings' will depend at the very least on the hardware, but also on what is running.
void flume
my voice changer wont work
Which voice changer? There are multiple. This is general AI server, not one for a specific tool.
Try filling out the template; it will help helpers figure out how to help you properly.
plucky crown
Some more details
It doesnt sound near to the target voice it should be adapting to my voice and just change the tone and pitch to replicate it exactly but its not doing that
It just sounds like my voice but the pitch is lowered and not a completely different person
modern sleet
Which voice changer? There are multiple. This is general AI server, not one for ...
can i send u ss
fringe heron
It just sounds like my voice but the pitch is lowered and not a completely diffe...
Target is a guy?
plucky crown
Yes