#✨│ai-help
brave garnetBOT
⠀
Settings for AMD GPUs 
Don't forget that your models needs to be converted in ONNX!
F0 Det.: rmvpe_onnx (suggested for all series)
7xxx XT cards: 112-128 chunk | +16384 extra
6xxx XT cards: 128-192 chunk | +16384 extra
5xxx XT cards: 192-256 chunk | +8192 extra
RX 580: 192-256 chunk | +8192 extra
RX 570: 192-256 chunk | +8192 extra
RX 560: 256-384 chunk | +8192 extra
Advanced Settings
Protocol : Sio or Rest
Crossfade: 4096 start 0.2 end 0.8
Trancate: 300
Silencefront: Off
Protect: 0.5
RVC Quality: Low
⠀
Made with love by Ilaria 💖
sand iris
windows username contains spaces?
proud elbow
@violet pecan use these settings
sand iris
ok it shouldn't be admin but whatever
brittle wing
ok it shouldn't be admin but whatever
do you know where i can get good valorant rvc models ?
sand iris
well its default
Honestly I don't know about that
People had that happen for thousands of different reasons
brittle wing
what means epochs ??
opal mica
Honestly I don't know about that
any other voice app?
lean burrow
feature extraction take 30 min +? 3 hour dataset, ilaria rvc
still doing feature extraction
['infer/modules/train/extract/extract_f0_rmvpe.py', '2', '1', '0', '/workspace/h/logs/DamYui', 'True']
todo-f0-2254
f0ing,now-0,all-2254,-/workspace/h/logs/DamYui/1_16k_wavs/0_1.wav
['infer/modules/train/extract/extract_f0_rmvpe.py', '2', '0', '0', '/workspace/h/logs/DamYui', 'True']
todo-f0-2255
f0ing,now-0,all-2255,-/workspace/h/logs/DamYui/1_16k_wavs/0_0.wav
f0ing,now-451,all-2255,-/workspace/h/logs/DamYui/1_16k_wavs/1_449.wav
f0ing,now-450,all-2254,-/workspace/h/logs/DamYui/1_16k_wavs/1_448.wav
f0ing,now-900,all-2254,-/workspace/h/logs/DamYui/1_16k_wavs/2_658.wav
f0ing,now-902,all-2255,-/workspace/h/logs/DamYui/1_16k_wavs/2_660.wav
f0ing,now-1353,all-2255,-/workspace/h/logs/DamYui/1_16k_wavs/4_276.wav
f0ing,now-1350,all-2254,-/workspace/h/logs/DamYui/1_16k_wavs/4_271.wav
f0ing,now-1804,all-2255,-/workspace/h/logs/DamYui/1_16k_wavs/5_488.wav
f0ing,now-1800,all-2254,-/workspace/h/logs/DamYui/1_16k_wavs/5_481.wav
f0ing,now-2250,all-2254,-/workspace/h/logs/DamYui/1_16k_wavs/6_92.wav
['infer/modules/train/extract_feature_print.py', 'cuda:0', '1', '0', '0', '/workspace/h/logs/DamYui', 'v2']
/workspace/h/logs/DamYui
load model(s) from assets/hubert/hubert_base.pt
move model to cuda
all-feature-4509
now-4509,all-0,0_0.wav,(149, 768)
now-4509,all-450,0_503.wav,(149, 768)
now-4509,all-900,1_447.wav,(149, 768)
now-4509,all-1350,2_250.wav,(149, 768)
now-4509,all-1800,2_657.wav,(149, 768)
now-4509,all-2250,3_464.wav,(149, 768)
now-4509,all-2700,4_270.wav,(149, 768)
brave garnetBOT
⠀
Settings for Nvidia GPUs 
F0 Det.: rmvpe (suggested for all series)
RTX 40-series: 80-96 chunk | +16384 extra
RTX 30-series: 96-112 chunk | +16384 extra
RTX 20-series: 112-128 chunk | +16384 extra
GTX 16-series: 128-192 chunk | +8192 extra
GTX 10-series: 128-192 chunk | +8192 extra
Advanced Settings
Protocol : Sio or Rest
Crossfade: 4096 start 0.2 end 0.8
Trancate: 300
Silencefront: Off
Protect: 0.5
RVC Quality: Low
⠀
Made with love by Ilaria 💖
strong mauve
helloo, i'm currently trying to test the ai voice changer and for some reason when i use the RVC types of voice changer, it can't pick up my voice. i've tested one of the sample models and for some reason the Beatrice one can pick it up just fine?
violet pecan
I mean where is it located
rare gobletBOT
Ayo? @violet pecan level 2 !!! 
pastel oak
I mean where is it located
on the left side above audio inputs
violet pecan
on the left side above audio inputs
i dont see it
pastel oak
i dont see it
latent kettle
for compiled version of applio what i have to download .zip or .exe??
brittle wing
what means epochs ?
latent kettle
what means epochs ?
epoch means a training cycle. that how much time it looked over your dataset
example: reading a book for 1 time is equal to 1 epoch
brittle wing
epoch means a training cycle. that how much time it looked over your dataset
when i use my voicechanger it sounds robotic do you know what i can do there ?
latent kettle
when i use my voicechanger it sounds robotic do you know what i can do there ?
yes. change your model. Czz it seems that it is not fully trained. for a good model 250-350 epochs are recommended and it has only 60
brittle wing
yes. change your model. Czz it seems that it is not fully trained. for a good mo...
do you know where i can get good valorant models ?
latent kettle
do you know where i can get good valorant models ?
brittle wing
okay
brittle wing
wdym ??
latent kettle
pastel oak
@latent kettle hes been asking the same thing over and over again, with robotic he just means that the voice model he uses is not made for realtime but rvc instead and theres no fix to that
to save you some time xd ilaria already helped with settings
pallid summit
sorry for the stupid question, i did a good training with the web-ui. i want to close the app and browser and record some more material for inferencing. when i run the app and open a new web-ui, will my model be there, or does it get deleted when everything is closed?
latent kettle
<@1174561195102056459> hes been asking the same thing over and over again, with ...
yes i asked him to change model czz it is traind on inly 60epochs
latent kettle
sorry for the stupid question, i did a good training with the web-ui. i want to ...
you can Resume the trainung if you have g and d files
pallid summit
you can Resume the trainung if you have g and d files
ok, training is done, so i'm guessing i'm safe to close and do inferencing later?
latent kettle
ok, training is done, so i'm guessing i'm safe to close and do inferencing later...
yes you can. you can find your model in Weights
pallid summit
thanks very much
rare gobletBOT
Ayo? @pallid summit level 1 !!!
pallid summit
?
pallid summit
oh, cool
is it really just one little 60mb pth file after all that training? it worked very well
proud elbow
is it really just one little 60mb pth file after all that training? it worked ve...
there are also several pth files in weights folder, would better test and pick which is best according to tensorboard and inference testing
pallid summit
interesting, thank you
pastel oak
idk what to do
https://docs.aihub.wtf/ "How to make AI Cover"
formal void
https://docs.aihub.wtf/ "How to make AI Cover"
thank u
brittle wing
what is the best tune for your free latina mommy voice ?
pastel oak
what is the best tune for your free latina mommy voice ?
it depends on your voice, trial and error
go to 12 and then adjust lower or higher
you have to wait until someone reviews it
rare gobletBOT
Ayo? @pastel oak level 23 !!!
minor wigeon
uhh i have problem with voice changer it says Pipeline is not initialized and then waiting generate pipeline
dense moon
uh how am i supposed to know how much the frequencies are at?
im chcking my dataset sample rate
the example showed the data beside the spectrogram, but this didnt xd
pastel oak
uhh i have problem with voice changer it says Pipeline is not initialized and th...
Try the 2nd option (deleting pretrain and model_dir) first, if that didnt solve it try the others
brittle wing
Try the 2nd option (deleting pretrain and model_dir) first, if that didnt solve ...
can you send the download link for amd ??
pastel oak
can you send the download link for amd ??
minor wigeon
Try the 2nd option (deleting pretrain and model_dir) first, if that didnt solve ...
well now it works but.. my voice cracks when im talking like im saying smth and then it says half of my word and stops
pastel oak
well now it works but.. my voice cracks when im talking like im saying smth and ...
Whats your GPU name?
What F0, chunk and extra did you select?
minor wigeon
Whats your GPU name?
What F0, chunk and extra did you select?
man i dunno i just want voice changer for real
pastel oak
man i dunno i just want voice changer for real
You just have to read out your settings..?
spam a few messages until youre lvl1 on the server to send screenshots
minor wigeon
You just have to read out your settings..?
this gpu?
minor wigeon
yes ok
send screenshot of voice changer
pastel oak
Ok it looks good enough
Audio Input: Microphone
Audio Output: Line 1 or CABLE Input (can you also tell me which one you have?)
Monitor: if you want to hear yourself, speakers/headphones
Can you click on "advanced settings" on the bottom and send screenshot of that too?
pastel oak
start 0.2 end 0.8
trancate 300
silencefront off
minor wigeon
okay lemme try
pastel oak
after that, do this (this is important)
minor wigeon
okay i try
near raft
Hello been testing out tortoise tts for a while and kind of a noob, where do i put the downloaded voice models to make them pop up in the list of models in the tortoise gui? sorry if this was the wrong forum for that question
dense moon
my voice sample rate is at 44.1, so do i choose the 40k or the 48k option? (using rvc disconnected btw)
pastel oak
my voice sample rate is at 44.1, so do i choose the 40k or the 48k option? (usin...
40k
azure marshBOT
RVC Guides (How to Make AI Cover)
brittle wing
# RVC Guides (How to Make AI Cover)
Started doing the english tutorial for making an ai voice model and I'm at the stage of setting up TensorBoard. The tutorial is showing the page to have a search bar which was used to get a graph up like in the image below, Not sure why but I don't have one it's just empty space, Why's this and how can I fix it? Thanks in adv
Nvm not in the image below I just realised you can't link images here XD
pastel oak
Nvm not in the image below I just realised you can't link images here XD
You need to spam a few messages to get lvl1 on the server to send pics
pastel oak
Started doing the english tutorial for making an ai voice model and I'm at the s...
At which part did you fail? can you like point me at the right direction on the guide
did you fail here or at a latter stage?
Or is the graph empty?
brittle wing
You need to spam a few messages to get lvl1 on the server to send pics
Ah makes sense
brittle wing
did you fail here or at a latter stage?
Was past that. Down where it says "In the search bar, type "g/total". This will be the graph you'll monitor."
pastel oak
Was past that. Down where it says "In the search bar, type "g/total". This will ...
Yes so if you havent trained the model yet its not going to show you anything
it will show up after you start training
itll take a few minutes potentially
brittle wing
No but my problem isn't the graph is empty, It's that the search bar they refer to doesn't seem to be on my page like theirs, So I can't actually get the graph up. I think it may be a slightly newer version that works a bit different as there were some other differences I noticed.
gentle inlet
What even is all this stuffz tbh 😔
graceful obsidian
uh how am i supposed to know how much the frequencies are at?
It's simple
Red is more or less where stable frequency response range ends
Purple is where it's unstable, lossy, damaged, semi-retrieved
Dark blue is null, no vocal-data, frequencies.
pastel oak
No but my problem isn't the graph is empty, It's that the search bar they refer ...
Uhh, it shouldnt be any different
you can also manually find the graph
pallid cedar
Hello. I have a question
Why is the model I send to Gradio downloaded normally but I can't select the model name?
I updated and everything, but it doesn't change
junior glade
yo
i was trying to train rvc with colab and got Error preprocessing data...
previously i`d fixed that installing numpy 1.23.5 but now this and later versions don't work
can you recommend me how to solve this problem or what other colab i can try to train models?
brittle wing
Do I have to split the audio for the dataset into 15 seconds (the old method) or there's no need I can use a whole full Acapella of the singer?
rare gobletBOT
Ayo? @brittle wing level 5 !!!
brittle wing
can you help me wit applio
brittle wing
can you help me wit applio
What's your issue
wispy burrow
You didn't select a voice model and/or voice sample (audio) to inference.
RVC-GUI is quite old and outdated though.
brittle wing
Uhh, it shouldnt be any different
Oh thats odd 😭
rare gobletBOT
Ayo? @brittle wing level 1 !!!
brittle wing
you can also manually find the graph
Where would I find it? I tried looking and couldn’t find it for the life of me
brittle wing
You didn't select a voice model and/or voice sample (audio) to inference.
RVC-GU...
Just a quick question.I don't have to split the dataset into 15 second samples right?
That's the modern method?
Applio also has voice model blend option;)
proper shale
Just a quick question.I don't have to split the dataset into 15 second samples r...
yeah no need
proper shale
This wasn't needed for a long time tho
wispy burrow
Even better.
RVC splits your dataset into 1-5 second clips when you do "Process Data".
wispy burrow
Looks like W-Okada.
#🔍│help-w-okada
proper shale
yeah, it is
dunno what's the issue tho
wraith fiber
Looks like W-Okada.
<#1159290161683767298>
i nedd ask in okada ?
rare gobletBOT
Ayo? @wraith fiber level 2 !!!
proper shale
might be spec related?
brittle wing
This wasn't needed for a long time tho
Last year in summer?
Whats the most realistic female voice file out there im voice trolling my friend :>
brittle wing
If you were using Sovits... maybe it was needed
Crepe RVC
wispy burrow
i nedd ask in okada ?
You might get a response/solution faster if you post there.
echo pelican
guys, i got a problem with the audio
hybrid flax
DTYPE
thick cloak
tensor board does not work for me
the graphcs areng showing
nevermind
got it
so whenever this graph starts going back up is when its overtraining right
terse wave
а где скачивать голоса
thick cloak
а где скачивать голоса
proper shale
so whenever this graph starts going back up is when its overtraining right
not necessarily, it's a bunch of factors
go to scalars tab and pin the following graphs:
d/total, g/total, g/mel and g/kl
if g/mel and g/kl start increasing alongside g/total, yeah that's overtraining
thick cloak
how can i pin a graph
proper shale
how can i pin a graph
i think there's a pin button when u search for a graph?
or u can just search for each one whenever u wanna see the graphs
daring raft
is there any better option than Replay for interference for MAC users?
And is there an online option where u can merge models?
thick cloak
is it overtraining?
proper shale
is it overtraining?
nope, g/total seems to still be going down
thick cloak
nope, g/total seems to still be going down
alr
humble furnace
What should I do if a woman's voice lisps and there are words?
hard crystal
hi, so i'm wondering if there is a 48k version of RIN_E3 or OV2super for RVC V2 custom pretrains, and i am wondering if 48k has any benefits over 40k for RVC models?
hasty wren
hello, I'm new here and I wanted to see how all of this works and when I try to create AI cover with google collab I'm stumbling upon this error in step 2:
"JSONDecodeError"
hard crystal
"JSONDecodeError"
for the credentials?
rare gobletBOT
Ayo? @hard crystal level 1 !!!
hasty wren
File exists, proceed with creation of creds and client
hard crystal
ive had that happen before, i just removed all the google auth code from the downloader
proper shale
hi, so i'm wondering if there is a 48k version of RIN_E3 or OV2super for RVC V2 ...
There isn't any benefit to use 48k as it has more noise problems
proper shale
"JSONDecodeError"
what colab notebook are ya using
hard crystal
There isn't any benefit to use 48k as it has more noise problems
oh, so if i was to make a 40k that was able to reach better pitches, would i just make the dataset have higher and lower pitches?
like if i wanted it to sound more accurate, would i kind of spread-out samples of different ranges
hasty wren
ive had that happen before, i just removed all the google auth code from the dow...
woah, not sure I'm advanced enough to know where to look for it
hard crystal
where are you downloading the model from?
hasty wren
well, I just copy the link from this discord
proper shale
like if i wanted it to sound more accurate, would i kind of spread-out samples o...
yeah give the dataset some range, it'll give the model some range
btw you should use 32k always imo
hard crystal
oh how come?
proper shale
oh how come?
so it learns how the person does that pitch and can replicate that when needed?
proper shale
https://colab.research.google.com/drive/1Gj6UTf2gicndUW_tVheVhTXIIYpFTYc7?usp=sh...
this is outdated
proper shale
so 40k is lower pitch accuracy?
nope not necessarily
hard crystal
wait it is?
proper shale
wait i got confused
so
32k, 40k and 48k are sample rates, aka resolution of ur files
hard crystal
wait it is?
ive used this colab all the time, it stopped working a month ago for me, even after modifying it abit
proper shale
no matter which sample rate you pick, it'll be able to replicate the pitch u give it
proper shale
ive used this colab all the time, it stopped working a month ago for me, even a...
yeah you should use the newer stuff
-colab
proper shale
applio, ilaria rvc
proper shale
maybe even try Ilaria rvc mainline which is in beta
hard crystal
so overall
should i use 32k or 40k?
im exporting my samples from a software that has sample rates up to* 96k
night fern
raise LocalProtocolError("Too little data for declared Content-Length") h11._util.LocalProtocolError: Too little data for declared Content-Length
What does this mean/
sterile crane
does anyone know what's the best way to speed up training when training locally
proper shale
should i use 32k or 40k?
32k
proper shale
` raise LocalProtocolError("Too little data for declared Content-Length")
h11...
huh, never seen this before
anything else before that? and what fork are you using?
night fern
anything else before that? and what fork are you using?
applio precompiled
hard crystal
does anyone know what's the best way to speed up training when training locally
you could increase batch size, and cache datasets
rare gobletBOT
Ayo? @hard crystal level 2 !!!
proper shale
does anyone know what's the best way to speed up training when training locally
Make sure no other background tasks are taking away the GPU power - also, in case you haven't started training yet you can mess with the batch size and what @hard crystal cited above
proper shale
caching datasets is only really useful for smaller datasets, but it makes a difference
just make sure ya have enough VRAM for everything
sterile crane
you could increase batch size, and cache datasets
thanks
sterile crane
Make sure no other background tasks are taking away the GPU power - also, in cas...
thanks
proper shale
Hmmm, that's pretty weird... frankly don't know what could be causing that
night fern
Hmmm, that's pretty weird... frankly don't know what could be causing that
It's when inferencing a song
it still works but sounds robotic
hard crystal
It's when inferencing a song
this error has a ticket for it in both:
the fastapi repo: https://github.com/tiangolo/fastapi/discussions/7187
the uvicorn repo: https://github.com/encode/uvicorn/issues/841
but i havent actually seen a ticket for it on the appilo github one
you could submit it if you wanted to
but IMO i think theres a problem with one of the projects dependencies
calm mauve
Where do I find the voice samples need to load for RVC? I can find some character and index files, but never the samples. I'm new to this; what am I missing?
rare gobletBOT
Ayo? @calm mauve level 1 !!!
thick cloak
i cant find the index file for my model
wispy burrow
Where do I find the voice samples need to load for RVC? I can find some characte...
There are no voice samples.
Just ModelName.pth & added_IVF*.INDEX
thick cloak
i cant find the index file for the model i just trained
whats the index file needed for
celest sorrel
quick question, what's the difference between steps and epoch? what does the number mean in each?
rare gobletBOT
Ayo? @celest sorrel level 1 !!!
compact elbow
calm mauve
There are no voice samples.
Just ModelName.pth & added_IVF*.INDEX
How are you making the output sound like the character with these two files only? The only option I see for VOICE, (for example using Tortoise AI Voice Cloning application) is RANDOM or my own microphone recording, which totally changes the voice outputted, so why do I need the .pth and index file if it ends up being a random audio output anyway?
I have a lot to learn obviously, and I'm sorry to ask a stupid question that is no doubt easy for almost everyone else.
I'm confused on how to get the character's voice without audio samples being involved... don't I need to have something other than random or microphone in the AUDIO dropdown PER character?
wispy burrow
Oh, it's TTS.
Yeah, I think those do need samples...
calm mauve
Ahhh ya I just noticed some samples with the #voice-model submissions
Where do I place those? Do you know?
split dome
hi guys. howto find german voice?
junior crane
Hey, so I am having a problem like activating the voice profiles? There is one that works by default idk why this one won't.
wispy burrow
Ahhh ya I just noticed some samples with the #voice-model submissions
Where do...
I think those are demos for us to audition the voice Model.
Also RVC Models are not TTS Models.
split dome
guys which software do u use for voice change?
rare gobletBOT
Ayo? @split dome level 1 !!!
proper shale
guys which software do u use for voice change?
depends on what u wanna do, for realtime we use w-okada
split dome
yes for realtime live voice change
calm mauve
I think those are demos for us to audition the voice Model.
Also RVC Models are ...
hmmmm, I'm ultimately desiring a solution that lets me type any line into it, and have it spoken back to me AS a specific character. Maybe I'm choosing the wrong solution?
Do you know of a tutorial to help me accomplish this? I don't want to keep bugging you or others here if I'm using the wrong solution for this...it seems like something that I was using was designed for this, but maybe not?
brave garnetBOT
⠀
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 💖
split dome
Thank you 🙂
proper shale
Thank you 🙂
You're welcome :)
split dome
i downloaded first voice . ai but i think it is scam, cause i need to pay for coins there. so i will delete it now. and install w okada
proper shale
i downloaded first voice . ai but i think it is scam, cause i need to pay for co...
well yeah voice.ai is paid and honestly not great compared to w-okada
open source ftw
misty stone
great voice changer but my pc cant handel it LOL
proper shale
great voice changer but my pc cant handel it LOL
yeah it's sadge
glass cypress
hey I'm getting this error trying to open Ilaria
proper shale
but if it works it works well
wispy burrow
hmmmm, I'm ultimately desiring a solution that lets me type any line into it, an...
RVC Models are used to change the voice from audio file or to one's voice in realtime..
TTS Models are different.
I am not familiar with them though.
proper shale
hey I'm getting this error trying to open Ilaria
hmmmmmm
misty stone
yeah it's sadge
ill just test this
brittle wing
Please make a model robot maksim
proper shale
maybe try removing the spaces in the folder name? don't think that's the issue tho...
split dome
"Virtual Audio Cables (mandatory)" do i need this also for w-okada?
proper shale
"Virtual Audio Cables (mandatory)" do i need this also for w-okada?
if ya wanna use the changed voice on any other application, yes
hard crystal
hey so, what would be the best time for each dataset sample? using the rvc-svc-dataset-maker
it says on the github 10 seconds split but i have 300mb worth of training data
so should it be longer?
proper shale
hey so, what would be the best time for each dataset sample? using the rvc-svc-d...
you don't need the rvc-svc-dataset-maker
splitting is no longer necessary
;)
glass cypress
Which is better, applio or Ilaria?
hard crystal
or i can just run it all through rvc and let it preprocess right
dense moon
It's simple
uh when comparing to the example shown in the doc?
proper shale
would it matter if there is silences?
yeah, silences screw up training
graceful obsidian
yours non readable.
stuff's simple.
any - 37~khz = 32khz model
37 - sub 38-39khz or anywhere close to 42/43khz = 40k model
for advanced users and audio savvy people;
38+ khz audio applicable for 48khz models
graceful obsidian
uh when comparing to the example shown in the doc?
^
you look for avg, not peaks
either way, model you go for should, for simplicity, be around what your audio's avg frequency response is
now that's all i can say. Not up for helping anyone rn + late at night
dense moon
Aight, well its training rn, just hope its good enough
hard crystal
is there any good spectrograms to find the audio frequencies?
fleet valve
https://www.youtube.com/watch?v=w1EBzQ_dF_M can anyone help me find this voice @opaque shoal
split dome
which file type do i need to upload in w-okada
proper shale
is there any good spectrograms to find the audio frequencies?
spek's a good one
proper shale
which file type do i need to upload in w-okada
.pth is the model, ya dont need index
hard crystal
spek's a good one
yeah i downloaded it lol
what should i go for based of this?
i hope the data is legible
split dome
i downloaded w-okada and uploaded a pth girl voice. but it sounds cringe for me like a robot lol.
quartz coral
I wanna train voice for cod youtuber merkmusic but dunno how much of audio in mins do i need alsk on mangio rvc i do see train yes but how to do it any help?
rare gobletBOT
Ayo? @quartz coral level 1 !!!
proper shale
i downloaded w-okada and uploaded a pth girl voice. but it sounds cringe for me ...
change tune n stuff
see what works
proper shale
I wanna train voice for cod youtuber merkmusic but dunno how much of audio in mi...
like, 5-10 mins?
tutorial for you
slender blaze
Did you manage to make it work on mac ? I'm in teh same boat
rare gobletBOT
Ayo? @slender blaze level 1 !!!
quartz coral
like, 5-10 mins?
Thanks i have 7-ish minutes of voice but i will try to have 3 more
hard crystal
@proper shale what sample rate should i use based on this spectrographic output?
humble minnow
i havent used the site in a few months, does it not work anymore?
proper shale
i havent used the site in a few months, does it not work anymore?
yeah this is old already
hard crystal
i havent used the site in a few months, does it not work anymore?
thats the old colab
proper shale
-colab
proper shale
use the newer ones like ilaria rvc and applio
proper shale
<@622856015444049937> what sample rate should i use based on this spectrographic...
always use 32k imo
hard crystal
what about quality?
does it effect audio quality
like compared to 40k, would 32k be worse in any regards?
proper shale
like compared to 40k, would 32k be worse in any regards?
i mean, not much
it's better 99% of the time anyway
quartz coral
like, 5-10 mins?
I read entire tutorial and i have three question
1.are datasets like audios for model?
2.What is epoch i searched up and still dont inderstand amd what is reccomend epoch to usw
3.Do i need to install tensorboard before training the model
proper shale
I read entire tutorial and i have three question
1.are datasets like audios for...
yes, datasets are audios for model
epochs are the cycles of training, there isnt a recommended amount since it varies per model
yes, you need tensorboard to help u know when its overtraining
quartz coral
Also one more thing can i use for only singing for traiing the model or it has to be speech causwe for example robbie williams sings mostlty not talk
quartz coral
yes, datasets are audios for model
epochs are the cycles of training, there isnt...
Thank you
pastel oak
I read entire tutorial and i have three question
1.are datasets like audios for...
adding to it if it isnt clear, tensorboard kinda takes away the need for a "Recommended epoch", it shows you when youve overtrained which can be at 250 epochs, could be at 400 etc.
proper shale
Also one more thing can i use for only singing for traiing the model or it has t...
yes u can use for both
marble plaza
What’s the best website to make ai song covers?
pastel oak
What’s the best website to make ai song covers?
On AI Hub we offer a local way of doing it and an online/colab/cloud way
everything you need is on the guide https://docs.aihub.wtf/essentials/how-to-make-ai-cover/
quartz coral
adding to it if it isnt clear, tensorboard kinda takes away the need for a "Reco...
Thanks
quartz coral
yes u can use for both
I forgot to ask since i can use singing does it have to be vocals only or there can be istrumentals?
proper shale
I forgot to ask since i can use singing does it have to be vocals only or there ...
vocals only yeah
azure marshBOT
RVC Guides (How to Make AI Cover)
split dome
how can i train a mp3 to a pth file?
proper shale
how can i train a mp3 to a pth file?
split dome
there is no one click tool? online tool
proper shale
there is no one click tool? online tool
yeah no "one click stuff"
split dome
ok thx
rare gobletBOT
Ayo? @split dome level 2 !!!
proper shale
but there are online tools
valid veldt
Having some issues getting the AI audio to come through my in game mic Specifically Helldivers 2.
I changed the ai voice output to a virtaul cable connector and thusly made set the cable as the input for the in game mic through my windows settings. This works with my discord app but when I try to do it with helldivers it still takes in The normal input of my mic then not the AI voice.
Sorry if that was a shit explanation I'm just really confused cuz I followed the directions I found online to a T
pastel oak
Having some issues getting the AI audio to come through my in game mic Specific...
Does Helldivers 2 not have an option to manually select mic?
valid veldt
Nope the games has a very simple audio settings the only way I found online to change it is through my computer audio mixer settings.
pastel oak
And with windows settings you mean sound device (where you select playback/recording etc.) you selected CABLE Output as your primary device?
valid veldt
Yes
valid veldt
Let me try that real quick
pastel oak
If that didnt work
Select the cable as your "default COMMUNICATION device". This seemed to have helped a lot of people
so select something als as your default, and select cable as communication i think
solemn musk
im using it rn and like it has a long delay and it keeps cutting at one part for some seconds then looping the problem again
ocean warren
where do i find the disconnect prevention method
valid veldt
Seems to have work 👍👍
rare gobletBOT
Ayo? @valid veldt level 1 !!!
valid veldt
Big thanks! You have saved democracy!
Time for Peter Griffin to mess with some randoms.
pastel oak
im using it rn and like it has a long delay and it keeps cutting at one part for...
are you talking about the voice changer?
whats your gpu?
pastel oak
well a cpu is not strong enough to run it, you dont have a gpu then?
quartz coral
How to know which sample rate to select for my voice to train?
rare gobletBOT
Ayo? @quartz coral level 2 !!!
pastel oak
theres an online colab version that uses your internet. you can try that outo if you want
solemn musk
ale
rare gobletBOT
Ayo? @solemn musk level 1 !!!
solemn musk
alr
quartz coral
I am going with robbie williams but just curiius how do i know which of tjese options i select?
solemn musk
thanks
pastel oak
How to know which sample rate to select for my voice to train?
not sure if its working atm its not for me
tbh you cant go wrong with training it on 32k sample rate, its good in almost all cases
quartz coral
Ok will try with 32k
warm skiff
where do i find the google collab to use the ai voice
pastel oak
where do i find the google collab to use the ai voice
rvc (making ai covers) or voice changer
warm skiff
rvc (making ai covers) or voice changer
rvc
pastel oak
rvc
https://docs.aihub.wtf/rvc/cloud/ilaria-rvc/
i recommend ilaria or rvc disconnected
warm skiff
https://docs.aihub.wtf/rvc/cloud/ilaria-rvc/
i recommend ilaria or rvc disconne...
oh okay thank you
rare gobletBOT
Ayo? @warm skiff level 1 !!!
quartz coral
I have a problem with tensorVENV when i open normally is sags another program is currenlty using this file and when i open it with administaror it immedaily closses and there is no icon is this command prompt?
ocean warren
"MessageError: Error: credential propagation was unsuccessful" in rvc disconnected
what do i do
thin stump
how do i perfectly isolate vocals
On Mvsep, you can use a isolation model called BS Roformer Viperx
hazy thunder
is this local?
thin stump
It's usually not this busy but that's because it went down earlier. https://mvsep.com/en
hazy thunder
i dont wanna wait a lot
proper shale
is this local?
yeah no it's online
thin stump
i dont wanna wait a lot
You'll have to use UVR 5 then and use the MDX23 model https://ultimatevocalremover.com/. That's local use
proper shale
You'll have to use UVR 5 then and use the MDX23 model https://ultimatevocalremov...
i think it's possible to download the bsroformer models actually
in the download center
thin stump
i think it's possible to download the bsroformer models actually
It's an MVSEP exclusive apparently
quasi wigeon
ever time i paste an ai voice link an error pops up what do i do JSONDecodeError
brittle wing
and whats the best model for removing background noises?
Uvr denoise?
thin stump
and whats the best model for removing background noises?
Dereverb-Deecho at 30 aggression then you can UVR denoise at 50-100
proper shale
ever time i paste an ai voice link an error pops up what do i do JSONDecodeError
using outdated colab, use newer stuff
-colab
brittle wing
But for UVR noise
proper shale
applio, ilaria rvc
hazy thunder
Dereverb-Deecho at 30 aggression then you can UVR denoise at 50-100
ok ty
brittle wing
Harmonify colab
quasi wigeon
using outdated colab, use newer stuff
use a new colab link or another ai?
hazy thunder
how much does segment size really affect the output
proper shale
use a new colab link or another ai?
newer colab link yeah
quasi wigeon
could you paste one here
brittle wing
quartz coral
I have a problem with tensorVENV when i open normally is sags another program is...
What to do with this cause if tensorboard doesnt launcg or atlast this file then i really dont dont knownwhat to do
brittle wing
quasi wigeon
says access denied
rare gobletBOT
Ayo? @quasi wigeon level 1 !!!
brittle wing
says access denied
Ask for access.
warm notch
does anyone know if theres any new versions of RVC that came out recently
quartz coral
Ok am i gonna assume you cant help me ok i will to figurw ir out Somehow god help with tjis😅😂...
proper shale
could you paste one here
https://docs.aihub.wtf/rvc/cloud/ilaria-rvc/
is a guide too
proper shale
does anyone know if theres any new versions of RVC that came out recently
nope, nothing new in terms of ai features yet
brittle wing
says access denied
Nvm gotcha
AICoverGen WebUI -https://colab.research.google.com/github/hinabl/AICoverGen-Colab/blob/main/Hina_Mod_AICoverGen_colab.ipynb
Ai cover gen no webui - https://colab.research.google.com/drive/1u1brjK8IZt647UsbZuGYfW29oFM2I4tk?usp=sharing
RVC DISCONNECTED - https://colab.research.google.com/drive/1XIPCP9ken63S7M6b5ui1b36Cs17sP-NS#scrollTo=Q7JhsK5oYByc
EASYGUI - https://colab.research.google.com/drive/1r4IRL0UA7JEoZ0ZK8PKfMyTIBHKpyhcw
Applio Colab - https://colab.research.google.com/github/iahispano/applio/blob/master/assets/Applio.ipynb
Harmonify - https://colab.research.google.com/drive/1X8YR4Ruv7zzY8YAMPTfC7hkxqT_d4Q5d
Ilaria RVC - https://colab.research.google.com/drive/16LkwvFZeudTpUOsE_6bMjOq2qkxFo8Hr?usp=sharing
There they are
quasi wigeon
thankyouu
thin stump
how much does segment size really affect the output
Use 512 segment size. UVR hasn't been consistent with the description for window size until now
warm notch
also I have aorund 10 minutes of voice for my model (most of the time) when I train them and I train them at 100 epochs. Is this too low? Because I keep getting mixed opinions. Some people say its too little and some people say going up higher wouldn't make a difference
quartz coral
Do i open tensorVENV with run with adminastratlr or just a normal ope cause if normal opet tbis happns
warm notch
try refreshing ur windows explorer in task manager
warm notch
do u have nvidea gpu?
quartz coral
Maybe i should close chrome i dunno??😐
quartz coral
do u have nvidea gpu?
Where can i check that?
warm notch
damn ok nvm I was gonna say just download the one I have but it unfortunately only works with Nvidea GPU's
quartz coral
Fuuuuuckkk
rare gobletBOT
Ayo? @quartz coral level 3 !!!
quartz coral
So i cant dowload it?
warm notch
nahh
rare gobletBOT
Ayo? @warm notch level 3 !!!
quartz coral
Wdym?datasets?
pastel oak
So i cant dowload it?
i think theres an amd version
warm notch
did they made an amd version?
pastel oak
You'll have to use UVR 5 then and use the MDX23 model https://ultimatevocalremov...
This is crazy, I didnt know theres an MDX23 model that works THIS good. I literally did an ensemble of like 5 different things, which takes 3 minutes to finish and achieved decent results - but this is a game changer, and faster. Thank you for that
warm notch
This is crazy, I didnt know theres an MDX23 model that works THIS good. I litera...
is there a new version of UVR that came out? Im not sure which version I have
pastel oak
is there a new version of UVR that came out? Im not sure which version I have
its a model you can download in UVR
quartz coral
https://huggingface.co/lj1995/VoiceConversionWebUI/resolve/main/RVC1006AMD_Intel...
What is this 2,9 gb is this for tensorboard?
quartz coral
Thank
pastel oak
if you run tensorboard on that it should work
since the mangio thing you downloaded was for nvidia only (i think)
thin stump
This is crazy, I didnt know theres an MDX23 model that works THIS good. I litera...
I stopped using ensemble after it stripped away my audio quality but ye MDX23 is all you need really
crimson pivot
Hi , Who can make a RVC models?
I need help.
warm notch
Hi , Who can make a RVC models?
I need help.
gotchu bro what questions do u have
quartz coral
for amd
And i am stupid for asking i am not really an expert on pcs but what is amd
😅
pastel oak
And i am stupid for asking i am not really an expert on pcs but what is amd
you have an AMD gpu
warm notch
I stopped using ensemble after it stripped away my audio quality but ye MDX23 is...
I just downloaded MDX23 what differences does it have compared to regular MDX-Net
crimson pivot
gotchu bro what questions do u have
I want to make the own models but my RVC program in computer didn't work for me.
Do you have any sites?
rare gobletBOT
Ayo? @crimson pivot level 1 !!!
warm notch
I want to make the own models but my RVC program in computer didn't work for me....
what GPU do u have
crimson pivot
The app kits was good but It's not free
dense marsh
How do I create models on mobile
crimson pivot
what GPU do u have
Nvidia
pastel oak
I want to make the own models but my RVC program in computer didn't work for me....
You can use colab, which is like a cloud thing you do it online
dense marsh
How do I create models on mobile
Which colab do I use
crimson pivot
You can use colab, which is like a cloud thing you do it online
Can I make in this site?
pastel oak
Can I make in this site?
Yes
thin stump
I just downloaded MDX23 what differences does it have compared to regular MDX-Ne...
There's no static or humming noises. I think most model makers are shifting to BS Roformer Viperx on Mvsep since it can extract instrumentals better than MDX23C
pastel oak
read the guide for all the steps
warm notch
There's no static or humming noises. I think most model makers are shifting to B...
ohhh ok makes sense thanks ima try it out
does anyone have the link for RVC1006Nvidea for @crimson pivot
I cant find it
pastel oak
does anyone have the link for RVC1006Nvidea for <@874277942971211797>
he said he wants colab
i already linked it
warm notch
ohh shit okok
slim wadi
how do i export a merged model?
crimson pivot
read the guide for all the steps
I saw but it doesn't say train
quartz coral
he said he wants colab
Wait isnr collab to have to pay foe it?
pastel oak
Wait isnr collab to have to pay foe it?
no
crimson pivot
I saw but it doesn't say train
@pastel oak
pastel oak
<@965636299509882980>
i dont know what you mean when you say it doesnt train
you wrote that 2 minutes after i posted it so idk what you tried yet
crimson pivot
i dont know what you mean when you say it doesnt train
Train my voice model
pastel oak
the thing does train voice
did you open it
wait no i might have a brainlag
sorry
here you go 💀 on me
faint moon
real time voice changer time delay why ? any reason or any solution to fix that time delay
crimson pivot
https://docs.aihub.wtf/rvc/cloud/rvc-disconnected/
Thank you so much🫂
pastel oak
real time voice changer time delay why ? any reason or any solution to fix that ...
what gpu do you have
faint moon
rtx 2050
quartz coral
What is difftemc between normal rtvx and rtvc disconnected?
faint moon
send me a screenshot of voice changer in dms
accept friend request
pastel oak
accept friend request
there is no friends request + you can dm me without request
lol
proud elbow
Wait isnr collab to have to pay foe it?
colab has free T4 gpu for one session daily
crimson pivot
send me a screenshot of voice changer in dms
Did you work with app kits ai?
If yes, Do you know sites like this?
pastel oak
Did you work with app kits ai?
If yes, Do you know sites like this?
I dont know about that sorry
proud elbow
Did you work with app kits ai?
If yes, Do you know sites like this?
that site is paywalled and dogshit
crimson pivot
I dont know about that sorry
Np bro
rare gobletBOT
Ayo? @crimson pivot level 2 !!!
crimson pivot
that site is paywalled and dogshit
Yeah I know and I looking for find sites like this but free !!!
quartz coral
that site is paywalled and dogshit
Well hope i didnt get scammed🙄🙄 qlso can i even download tensboard cause i dont have nvidiagpu
proud elbow
Well hope i didnt get scammed🙄🙄 qlso can i even download tensboard cause i don...
-colab
dense marsh
What colabs can be used to make models?
quartz coral
And without ndvidiagpu
bronze ocean
Do you guys know how to fix colab step 2 error
quartz coral
Suggestions for <@554736263534608394>
Thanks i will some of these
proud elbow
And without ndvidiagpu
colab runs gpu in cloud (i.e. Nvidia tesla T4 for free account)
lime kraken
How to make a ai cover?
quartz coral
In youtube you find ir from thkis channel
gusty kestrelBOT
@lime kraken, I have found 21 results that match your search!
Requested by ralves126
quartz coral
@pastel oak so link which you sent how do i use it and how do i download tensorboard with it
pastel oak
<@965636299509882980> so link which you sent how do i use it and how do i downlo...
the rvc1006amd thing you just paste the tensorboard file in the folder, run it and it installs it
should work normally
quartz coral
the rvc1006amd thing you just paste the tensorboard file in the folder, run it a...
Well i am out luck i still get the same goddamn error if open with run as admidantor nothinf its go with open normally it says another program is using this file
pastel oak
not sure maybe restart pc
quartz coral
Ok will try that
Also can i rely on this guide from here https://youtu.be/tnfqIQ11Qek can i train with rvc 2 disconccedted tho?
slim wadi
i reinstalled it and it wont open? its asking me to open urls
quartz coral
not sure maybe restart pc
Yup thqt was the problem but i still dont know if sholoud stick with tensor board in mangio or these one you sent me cause it works formangio and one the one you sent me
now what
rare gobletBOT
Ayo? @quartz coral level 4 !!!
proud elbow
then open localhost:6006
pastel oak
it works youre good. manually open the localhost url if it doesnt open
and it doesnt matter if you use mangio or mainline. the one you have now is slightly faster either way according to others
jolly yew
what is the most realistic and softest e girl voice i can get? Im willing to pay a reasonable ammount of required
magic rivet
hi
idk what im doing wrong
JSONDecodeError Traceback (most recent call last)
<ipython-input-13-598fdab96131> in <cell line: 31>()
31 if os.path.exists(config_path):
32 # File exists, proceed with creation of creds and client
---> 33 creds = Credentials.from_service_account_file(config_path, scopes=scope)
34 client = gspread.authorize(creds)
35 else:
5 frames
/usr/lib/python3.10/json/decoder.py in raw_decode(self, s, idx)
353 obj, end = self.scan_once(s, idx)
354 except StopIteration as err:
--> 355 raise JSONDecodeError("Expecting value", s, err.value) from None
356 return obj, end
😦
i made the ai myself and uploaded it to mega idk what im missing
rare gobletBOT
Ayo? @full token level 1 !!!
full token
btw whats the best for quality harvest or rmvpe_gpu?
magic rivet
heya! can someone send me a link on how to upload a model zip file? I wanna post mine on huggingface like you guys do
rare gobletBOT
Ayo? @magic rivet level 1 !!!
full token
heya! can someone send me a link on how to upload a model zip file? I wanna post...
in the pinned guide
magic rivet
to whoever made that tutorial fucking thank you, nothing else on this god damn earth worked
i super appreciate it
full token
rvc training doesnt seem to be saving the pth at the weights folder? pls help
peak vault
Hey, when I try to log into weights.gg with either my discord or gmail, it doesn't seem to work. It gives a server error message. Is anyone else having this problem?
azure marshBOT
RVC Guides (How to Make AI Cover)
rare gobletBOT
Ayo? @brittle wing level 1 !!!
cyan ruin
why cant i hear myself in voice.ai
i was able to before i restarted my pc
is it a setting?
glad zealot
why cant i hear myself in voice.ai
Not supported here
For voice changer we use W-Okada voice changer #🔍│help-w-okada
cyan ruin
my bad
glad zealot
Basically how many times it's been trained
Not always but most of the time yes
Basically think of someone training to do something, if they only trained a few times, it's not really gonna change much, but if they train enough they will be great at it, if too much training it may cause their body to break
obtuse stream
can anyone tell me how the appio overload sensor works? because it's a little strange that i set the workout to 300 epochs, but the sensor can be set to a maximum of 100 and it stops the workout, but it doesn't matter what i set it to, it can be 5 or 10 and it still stops. it wouldn't make sense to set a number for the workout and then the sensor detects if there is an overload and stops it. it's totally unnecessary to set a number.
I don't think you should set a number for overtraining, but the program itself should detect when it are overtraining and stop himself.
urban viper
probs ran out of disk space on C drive; i keep getting that on other programs whenever that happens
if you still have plenty of disk space on C (>50GB), then perhaps the page file is being assigned manually and UVR tries to ask for more page file than windows can assign
start menu -> search "advanced system settings" -> click "settings" on the performance menu -> click on advanced tab -> virtual memory, click "change" -> check "automatically manage paging file size for all drives" which should make windows assign page file size depending on how much it needs rather than a fixed size
gusty kestrelBOT
@edgy anchor, I have found 1 results that match your search!
Requested by inotmarc
brittle jay
oh wait nvm
solid path
Stupid question, but are they any major version between these?
obtuse stream
does that mean that from 166 it no longer train your voice well?
obtuse stream
can anyone tell me how the appio overload sensor works? because it's a little st...
anyone??????
quartz coral
it works youre good. manually open the localhost url if it doesnt open
and it d...
Thanks but i got no dashboards are acti for current data set
pastel oak
beatrices are working but rvc voices are not working
what does the cmd say when you run other voices
sharp swan
[Voice Changer] Pipeline is not initialized.
[Voice Changer] Waiting generate pipeline...
[Voice Changer] Pipeline is not initialized.
[Voice Changer] Waiting generate pipeline...
[Voice Changer] Pipeline is not initialized.
[Voice Changer] Waiting generate pipeline...
[Voice Changer] Pipeline is not initialized.
[Voice Changer] Waiting generate pipeline...
its keeping
same thing
pastel oak
[Voice Changer] Pipeline is not initialized.
[Voice Changer] Waiting generate pi...
Try closing the app, deleting the “pretrain” and the “model_dir” folders and restarting it.
brittle wing
anybody have the applio rvc colab page
pastel oak
anybody have the applio rvc colab page
azure marshBOT
RVC Guides (How to Make AI Cover)
brittle wing
and does anybody have the ilaria rvc mainline colab
brittle wing
https://colab.research.google.com/drive/16LkwvFZeudTpUOsE_6bMjOq2qkxFo8Hr?usp=sh...
thanks
storm knot
Heya! Idk if this is for like..... asking for help with making the voices or not since I dont talk in here a lot, but I'm currently having an error when it comes to downloading an AI voice and idk if its simply on my end or if there's something wrong with the actual model-- If anyone can understand this, I'd def appreciate the help! I can usually understand whats going on when I look at an error message, but my brain is failing me today hdfjhfjdks-- Sorry if I'm just bein a lil stupid BEHE
JSONDecodeError Traceback (most recent call last)
<ipython-input-5-4619ab35a734> in <cell line: 31>()
31 if os.path.exists(config_path):
32 # File exists, proceed with creation of creds and client
---> 33 creds = Credentials.from_service_account_file(config_path, scopes=scope)
34 client = gspread.authorize(creds)
35 else:5 frames
/usr/lib/python3.10/json/decoder.py in raw_decode(self, s, idx)
353 obj, end = self.scan_once(s, idx)
354 except StopIteration as err:
--> 355 raise JSONDecodeError("Expecting value", s, err.value) from None
356 return obj, endJSONDecodeError: Expecting value: line 1 column 1 (char 0)
lapis oracle
Could you give me tips for the parameters in the song inference? I cannot get a good performance with female voices. pitch change to -7 works well with males but not females
storm knot
It honestly depends on what song you're using, but if making it lower doesn't work, I'd recommend raising it to a 5 or even 10 depending on the voice ^^
brittle citrus
voice delay
proven hill
i need help... in life you know...
rich ferry
Can you tell me how can I use it in game?
limpid tangle
Can you tell me how can I use it in game?
go to your audio settings in game and change the input to the cable thingy ig
signal mesa
how to get jetts vc?
willow bronze
Hello, someone know why I get that error?
proud elbow
Hello, someone know why I get that error?
what the console window says?
mild zephyr
i have a blank screen whenever i open the rvc
quartz coral
Can i train a voics rvc2 disconnected?
nimble garnet
can someone please help me?
brave garnetBOT
Made with love by Ilaria 💖
quartz coral
Also for my model of robbie williams could 500 epoch mi enough and for inactivity can i really click around so i dont get disconnected?
nimble garnet
can someone please help me?
lean agate
anyone knows any other rvc program for ai?
brittle wing
can someone tell me how I can open a tfevents file in tensorboard
forest sluice
can someone tell me how I can open a tfevents file in tensorboard
Put it into a directory, then tensorboard --logdir=path/to/tfeventsfolder
quartz coral
How do i know if overtrained the model?
latent kettle
How do i know if overtrained the model?
look at G and D loss on tensor board
also Kl and mel
if everything us goinig down you are good to go.
spring orchid
-colab
brittle wing
Put it into a directory, then `tensorboard --logdir=path/to/tfeventsfolder`
how do I install tensorboard locally to run this command?
quartz coral
What are those?Also when i open tensoardboard with that with that file why does it say no dashboardrds are active?i mean when i get to this step wait a min
When i do this step
forest sluice
how do I install tensorboard locally to run this command?
pip install tensorboard
lean agate
-colab
lean agate
-rvc
latent kettle
Will it no longer say no dashboarda are active?
are you training locally?
brittle wing
`pip install tensorboard`
'tensorboard' is not recognized as an internal or external command,
operable program or batch file.
rare gobletBOT
Ayo? @brittle wing level 11 !!!
quartz coral
are you training locally?
I will have to check in few hours cause i have to start training the model and i think i open the that file for tensorboard and then not sure also can use rvc 2 disconnected for training the nodel
knotty lintel
can someone help me make a voice model?
pastel oak
can someone help me make a voice model?
Do you mean the steps in how to create one or if someone else can literally create the model for you
rare gobletBOT
Ayo? @pastel oak level 25 !!!
knotty lintel
Do you mean the steps in how to create one or if someone else can literally crea...
either or, whatever works best
rare gobletBOT
Ayo? @knotty lintel level 1 !!!
acoustic sonnet
either or, whatever works best
go to the aihub docs, really comprehensive guide there
quartz coral
Do you mean the steps in how to create one or if someone else can literally crea...
I have one quwstion can i create voice models with rvc 2 disconecce
pastel oak
either or, whatever works best
pastel oak
I have one quwstion can i create voice models with rvc 2 disconecce
do you mean "RVC Disconnected"? Yes
proven hill
someone needs help?
gusty kestrelBOT
@violet pecan, I have found 8 results that match your search!
Requested by sillyfoo1s
@violet pecan, I have found 7 results that match your search!
Requested by sillyfoo1s
quartz coral
do you mean "RVC Disconnected"? Yes
Yes thanks
somber plover
How do I upload my own ai voice models?
brittle wing
some helppls
HttpError Traceback (most recent call last)
/usr/local/lib/python3.10/dist-packages/pydrive/files.py in _FilesInsert(self, param)
367 param['media_body'] = self._BuildMediaBody()
--> 368 metadata = self.auth.service.files().insert(**param).execute(
369 http=self.http)
5 frames
HttpError: <HttpError 404 when requesting https://www.googleapis.com/drive/v2/files?alt=json returned "File not found: 1UPB5nE8E_IQNYUVvpdHcf_Bym9OyGBUJ". Details: "[{'message': 'File not found: 1UPB5nE8E_IQNYUVvpdHcf_Bym9OyGBUJ', 'domain': 'global', 'reason': 'notFound', 'location': 'file', 'locationType': 'other'}]">
During handling of the above exception, another exception occurred:
ApiRequestError Traceback (most recent call last)
/usr/local/lib/python3.10/dist-packages/pydrive/files.py in _FilesInsert(self, param)
369 http=self.http)
370 except errors.HttpError as error:
--> 371 raise ApiRequestError(error)
372 else:
373 self.uploaded = True
ApiRequestError: <HttpError 404 when requesting https://www.googleapis.com/drive/v2/files?alt=json returned "File not found: 1UPB5nE8E_IQNYUVvpdHcf_Bym9OyGBUJ". Details: "[{'message': 'File not found: 1UPB5nE8E_IQNYUVvpdHcf_Bym9OyGBUJ', 'domain': 'global', 'reason': 'notFound', 'location': 'file', 'locationType': 'other'}]">
this is the problem im getting
willow bronze
what the console window says?
This is what my console say
low shard
HttpError Traceback (most recent call last)
/usr...
what google colab are you using?
low shard
How do I upload my own ai voice models?
check our docs https://docs.aihub.wtf
quartz coral
I need help before i start rvc disconnectwd since it says rvc or whatever doesnt suppoet rar orr 7 zip how to change it?
Is it somewhere on settings?
azure marshBOT
RVC Guides (How to Make AI Cover)
brittle creek
When I'm in step two, it's not going to add the voice model, since I clicked on all the plays before, what do I do?
quartz coral
Nvr i found it
amber zephyr
hey if use rin_e3 for training on a very clean datset would it be slightly better than training with original
anyone?
dense shore
how do u make ur own ai voice?
quartz coral
Since i am using rvc disscorned what is default crepe_hop_lenght?
does anzbodz know how to loadd dataset?
thin stump
Since i am using rvc disscorned what is default crepe_hop_lenght?
leave it as default
unless you are using mangio crepe. You'll never change it for rmvpe
quartz coral
i am using rvmpe
Also one more question load dateset cell is it this?
and how do i load my dataset from zip to see it here
willow bronze
Your local version 3.1.1 is the latest version.
And I am trying to do that
willow bronze
Your local version 3.1.1 is the latest version.
And I am trying to do that
When I click in convert it give me that error
quartz coral
also since i dont have ndvidia gpu can i set to like 500 later
willow bronze
Yeah I speak spanish
empty gazelle
is it possible to use a sovit file for real time voice changer, if so how can i upload it to the client
thin stump
Also one more question load dateset cell is it this?
I was afk, did you figure it out? You need to put the zip folder of the dataset inside the rvc disconnected folder of the google drive
quartz coral
I was afk, did you figure it out? You need to put the zip folder of the dataset ...
yup went to google drive and pasted robbie_williams but what now i am following step bz step from website rvc disconnected
rare gobletBOT
Ayo? @quartz coral level 5 !!!
thin stump
yup went to google drive and pasted robbie_williams but what now i am following ...
You have to make sure the folder name has no spaces or special characters like the underline you're using
rapid tinsel
thin stump
otherwise it gets confused
willow bronze
mandame un link de otro modelo que sepas que funcione y lo pruebo
Solo he probado con tts, no he indigado mas, ya que es el tts el que me gustaria usar hehehehe
thin stump
yup went to google drive and pasted robbie_williams but what now i am following ...
so just reupload the folder and you can load the dataset
quartz coral
otherwise it gets confused
oh shit i forgot but i got this
willow bronze
que va, me da error con todos
quartz coral
so just reupload the folder and you can load the dataset
also i got this and do i need to move to mangio rvc folder do i need to move to dataset folder there
grim bay
Is there a faster way to do this for free?
thin stump
also i got this and do i need to move to mangio rvc folder do i need to move to...
it takes a few minutes so refresh the google drive
page
brittle creek
?
rare gobletBOT
Ayo? @brittle creek level 2 !!!
willow bronze
@brittle wing Me recomiendas usar esta version, quizas es mas facil usarla para mi, quizas hay algo por dentro de la instalacion que hice mal, esta version se supone que te lo da ya todo bien configurado
grim bay
Idk how to use it on Colab
thin stump
also i got this and do i need to move to mangio rvc folder do i need to move to...
and it says experiment_name not defined so you missed a step
willow bronze
Perfecto, y como desistalo la que tengo? solo borro la carpeta y listo?
rare gobletBOT
Ayo? @willow bronze level 3 !!!
willow bronze
Perfecto, gracias
grim bay
Also, what should I do on my dataset when the UVR and the website are too slow?
Really, my laptop can't handle it
quartz coral
and it says experiment_name not defined so you missed a step
i mean i following steps from website i set all training varibables but wait do i need to click this cell before do this next one
grim bay
Don't ignoring me
quartz coral
Don't ignoring me
Just wait a bit i guess?
thin stump
i mean i following steps from website i set all training varibables but wait do ...
Click load datset then if you already set your training variables
grim bay
Just wait a bit i guess?
A bit means 9 hours on 5 minutes song
thin stump
A bit means 9 hours on 5 minutes song
mvsep is not that long, you can wait and do something else
grim bay
mvsep is not that long, you can wait and do something else
Well, how long you think it might takes?
thin stump
Well, how long you think it might takes?
You aren't logged in. That's why your queues are long
quartz coral
Click load datset then if you already set your training variables
now i got this error on previus steep i still dont undertand one thing do i need put robbie williams inside datasets on mangio
thin stump
now i got this error on previus steep i still dont undertand one thing do i need...
Run this step https://docs.aihub.wtf/rvc/cloud/rvc-disconnected/#2-set-the-environment. I don't see anything wrong here
grim bay
You aren't logged in. That's why your queues are long
So how long it might take for me to do about 1 dataset?
I'm not sure how long but there are 44 sentences = 44 .mp3 files right now
And it can be increase because that's not all of the dataset
quartz coral
Run this step https://docs.aihub.wtf/rvc/cloud/rvc-disconnected/#2-set-the-envir...
i did and it says 0s
thin stump
i did and it says 0s
Dm me?
thin stump
So how long it might take for me to do about 1 dataset?
I'm not sure how long bu...
You need to combine the audio together to reduce the amount of files. If you go above 100mb, it won't accept it
grim bay
You need to combine the audio together to reduce the amount of files. If you go ...
Well, that's suck if I have 100mb audio
Also which one is the best to separate vocals for AI cover? Not making model
thin stump
Also which one is the best to separate vocals for AI cover? Not making model
BS Roformer Viperx
hearty wyvern
@low shard like, I uploaded the voice, but how do I use it?
mortal quiver
Is there a link for colab
low shard
<@911742715019001897> like, I uploaded the voice, but how do I use it?
you made that voice right?
grim bay
Do we have UVR on Colab?
rare gobletBOT
Ayo? @grim bay level 10 !!!
low shard
Is there a link for colab
yep, check the cloud folder inside the rvc folder of our docs https://docs.aihub.wtf
hearty wyvern
you made that voice right?
I uploaded a model, I want to make a cover with Alastor's voice
rare gobletBOT
Ayo? @hearty wyvern level 1 !!!
low shard
Do we have UVR on Colab?
low shard
I uploaded a model, I want to make a cover with Alastor's voice
okay, do you got a good pc gpu?
hearty wyvern
okay, do you got a good pc gpu?
doesn't this work on mobile?? i'm not on pc rn, I'm having lunch🥲
brittle creek
I have already installed the model in voice but I can't click on the arrow to install it, what do I do?
low shard
doesn't this work on mobile?? i'm not on pc rn, I'm having lunch🥲
yup it does, then im gonna send you the online way, you can use one of these colabs:
please check them out and see whats the best for your needs, honestly i use ilaria rvc
brittle creek
I have already installed the model in voice but I can't click on the arrow to in...
?
silent vector
Can anyone help me with this?
ModuleNotFoundError Traceback (most recent call last)
<ipython-input-4-d9c4aae1a637> in <cell line: 4>()
2 #@markdown Link the URL path to the model (Mega, Drive, etc.) and start the code
3
----> 4 from mega import Mega
5 import os
6 import shutil
ModuleNotFoundError: No module named 'mega'
NOTE: If your import is failing due to a missing package, you can
manually install dependencies using either !pip or !apt.
To view examples of installing some common dependencies, click the
"Open Examples" button below.
All of the links I've used on the spreadsheet doesnt work
low shard
Can anyone help me with this?
ModuleNotFoundError Traceba...
what spreedsheet are you talking about and what google colab are you using
silent vector
low shard
I have already installed the model in voice but I can't click on the arrow to in...
did you click refresh and then clicked the arrow after it downloaded?
low shard
https://colab.research.google.com/drive/1Gj6UTf2gicndUW_tVheVhTXIIYpFTYc7?usp=sh...
both of those are outdated
low shard
Where can I find the new ones?
please check the colabs the cloud folder inside the rvc folder of our docs https://docs.aihub.wtf
and for searching voice models
you can search:
- #1175430844685484042
- #🔍│find-models
- https://weights.gg/
- https://huggingface.co/models (but watch out cus in hugging face there arent only rvc ai voice models)
- https://applio.org/models
if there isnt one, #1159289738314919936 or make it yourself with our docs guides https://docs.aihub.wtf
brittle creek
did you click refresh and then clicked the arrow after it downloaded?
Yes, but it doesn't give me the option to click on the screen, when I click it already clicks on the entire tab
silent vector
please check the colabs the cloud folder inside the rvc folder of our docs https...
Thanks man 👍
low shard
Thanks man 👍
yw
low shard
Yes, but it doesn't give me the option to click on the screen, when I click it a...
mmm maybe try on dekstop mode
or retry to zoom on the arrow and click on it
you should have a dropdown selection of the models you downloaded
silent vector
Is it alright if you help me find the rvc? its kinda confusing for me
ope nvm
found it
rare gobletBOT
Ayo? @silent vector level 1 !!!
silent vector
any rvc you recommend, or is it all pretty much the same?
low shard
any rvc you recommend, or is it all pretty much the same?
its not all the same they have pros and cons for each which you can read on their page, for making ai covers (also said as inference) i can recommend you ilaria rvc which is what i personally use, its still your choice tho
silent vector
alrighty, thank you
low shard
yw
mortal quiver
yep, check the cloud folder inside the rvc folder of our docs https://docs.aihub...
It doesn’t work for me
rare gobletBOT
Ayo? @mortal quiver level 1 !!!
mortal quiver
I got an error
low shard
The ones from the doc
there are like 5, which one are you using ?
north hazel
Hello
I'd like to use #1223583367077953629 message but i cant
find a link or something like that shown on a tutorial video i've seen! Could someone help please?
graceful obsidian
#🧬│ai-chat message
@modern marlin
Epochs do matter.
An epoch is more or less, in simple words; 1 run through your whole dataset. ( including gradient computations and so on under the hood )
willow bronze
Hello Why I get that error, someone know?
graceful obsidian
as a rule of thumb - Your training's dynamic is gonna dictate you when to stop ( at which epoch ) rather than " Imma train it for X epochs "
You can actively monitor the training using tensorboard.
It really depends you see, each case is unique
but typically, better the set, quicker the optimal solution
so, higher chances for quicker trainine / smaller required epochs count
as patterns are easier to spot and learn for gan
mortal quiver
there are like 5, which one are you using ?
I only seen two colab links in the doc and used both of them
graceful obsidian
it's really a matter of observing the loss tendency, dynamics
low shard
Hello
I'd like to use https://discord.com/channels/1159260121998827560/122358336...
he forgot, its fixed now, btw what yt tut are u watching
north hazel
@stable kiln hmmm why does discord tell me its not safe
rare gobletBOT
Ayo? @north hazel level 1 !!!
low shard
I only seen two colab links in the doc and used both of them
please just send the google colab you are using there are more
low shard
<@1158991133347684372> hmmm why does discord tell me its not safe
no thats just discord being discord, any download link on discord is being marked as not safe so yk to be sure when clicking on a right link, u can try urself to paste a download link on discord and click on it it gives u the same warning
graceful obsidian
For sure are, but in our case, being rvc
you mainly gotta be looking for few things
G and D ( total loss metric ) should be, in " idealistic " scenario, both going down
more or less, balanced
but that's not always the case in rvc
quartz coral
is s radeon rx 570 series good for apllio
stable kiln
<@1158991133347684372> hmmm why does discord tell me its not safe
Huh?
graceful obsidian
but that's not always the case in rvc
In fact, unstability or noisy graphs are common
low shard
https://www.youtube.com/watch?v=oOBjntI2xK0 !
oh thats outdated, its better if you follow the guides in our docs https://docs.aihub.wtf
graceful obsidian
by noisy I mean zigzaggy and those that go up keep the trending, then down, keep the trending
If you want more detailed information on such, go visit RVC Devs discord server and look through my past msgs
low shard
graceful obsidian
I used to explain lots of things related to rvc and tensorboard usage ( + overtraining and such spottings )
Cause current document at aihub isn't 100% correct or informative
and lacks in-detail explanations
unless they've changed it already
as I mentioned it last time I visited here
north hazel
hell nah thats too complicated
mortal quiver
please just send the google colab you are using there are more
Nvm I found a different one
low shard
hell nah thats too complicated
you are using an outdated version, you will probably run into problems which cant be fixed bc its outdated
graceful obsidian
Yet, the most important thing about rvc and tensorboard is syncing the logging interval
It is not synced on default. One has to do it manually and all " automated retrieval of steps " attempts should be ignored as they're innacurate and do not take into account rvc's internal segmenting and in essence, it can't be " estimated " or calculated with 100% precision
low shard
its better if you read our docs tho
north hazel
its better if you read our docs tho
i read them but i didnt understand 3. Convert the vocals
graceful obsidian
why?
hold on
::https://github.com/cy12345678/batch-file-for-tensorboard/
::check out my blog if you are interested, https://chunyangblog.wordpress.com@echo off
:: get hostname of your computer and save it to variable host.
FOR /F "usebackq" %%i IN (hostname) DO SET host=%%i:: use port 25565 as the tensorboard port.
set port=25565:: the link to the local tensorboard webpage is as follows
set address="http://%host%:%port%":: display the address in the command prompt
echo %address%:: ask user to key in the saved model directory, example "C:\tmp\mnist_model"
set /p UserInputPath=Key in model saved directory::: use default browser to open tensorboard webpage
explorer %address%::start tensorboard
tensorboard --logdir="%UserInputPath%" --port=25565 --bind_all:: to stop tensorboard, press Ctrl + C in command prompt, type "y" then hit enter.
pause
==== ^^^^^^^^^^^^^
make it a .bat
all you need aside that is:
pip install tensorboard
and call it a day
run the bat in any folder
in the cmd, put in the path to your tf events file
pretty straightforward
quartz coral
@low shardsince @thin stump told me in dms you areactive is radeon rx 570 series good for traiing voice in apllio
north hazel
what abt this message
graceful obsidian
generally 3.10.x
low shard
i read them but i didnt understand 3. Convert the vocals
it means if you got a good pc gpu check the local folder which will give you options, some have pros, some have cons, if you dont got a good pc check the cloud folder and you will find rvc with different pros and cons that you can use online based on your needs
but the rvc gui your using is very outdated
graceful obsidian
0 control over most important things + no control over what they do with your models
low shard
will prolly js run into errors
north hazel
but the rvc gui your using is very outdated
i didnt install anything
and sadly i wont
willow bronze
it means if you got a good pc gpu check the local folder which will give you opt...
can you please help me on that? THANKS #✨│ai-help message
north hazel
i'll leave later, probably
thin stump
<@911742715019001897>since <@1166084332205649945> told me in dms you areactive ...
I meant to ask that in general since it's active right now 
low shard
i didnt install anything
you can use the online rvc version in the cloud folder if ya want
graceful obsidian
training on rx 570 is not worth it at all
I am an user of such
4 or 8 gb is not enough for good models
in 70% of cases you won't fit the appropriate batch requirements
unless working with a small dataset
low shard
can you please help me on that? THANKS https://discord.com/channels/115926012199...
sorry but i dont really know much abt applio, maybe @half cove knows that error
graceful obsidian
cause torch, pytorch, cuda
quartz coral
I meant to ask that in general since it's active right now <:shrug:1159570387953...
whoops i thought it was help-rvc only ok ask and tell me back in dms
graceful obsidian
it's about vram
willow bronze
sorry but i dont really know much abt applio, maybe <@936673139419664414> knows ...
Ohh okey, no problem
graceful obsidian
I'd say, 8 is minimum comfortable but recommended minimum I'd say is 12 gig
whereas recommended overal is 15 or 16 gig
thin stump
whoops i thought it was help-rvc only ok ask and tell me back in dms
no, no. Not #🧬│ai-chat. Nevermind, you already got your answer
graceful obsidian
that grants you peace even with 30-40min sets
on batch 16 / 18 or even 20 sometimes
( in reality, you almost never go above 16 or 18
task manager, find your gpu in performance tab
willow bronze
@half cove can you please help me on that? THANKS #✨│ai-help message
graceful obsidian
also, training ( unless stuff's changed in non streamline rvc
on amd
won't work
or works but has some issues
ye
what's your gpu
16 then
that's more than enough
perfect even ~
last siren
why applio didnt work
graceful obsidian
it really depends on how good your ears are or should I say, how much you've trained them to spot troublesome models
sometimes it sounds " fine " but actually such models have lots of issues people aren't aware of
could be glitches, could be sibilants or higher-frequencies reconstruction issues etc.
last siren
-colab
low shard
why applio didnt work
!howtoask
patent trellis
low shard
i just found the ai i needed in kits.ai
srry but its simpler i'll use that
yea we arent affiliated with kits.ai bc its greedy and focus more on money than actual quality lmao, we use rvc which is js all free cus no one is paying over 50 dollars for ai covers
alpine juniper
Can someone help me with a JSONDecodeError
north hazel
yea we arent affiliated with kits.ai bc its greedy and focus more on money than ...
i actually didnt pay, when i tried to save it didnt let me but they forgot i can screen record 😎 (it worked lololo)
alpine juniper
Can someone help me with a JSONDecodeError
Please
rare gobletBOT
Ayo? @alpine juniper level 1 !!!
low shard
<@911742715019001897>since <@1166084332205649945> told me in dms you areactive ...
im areactive what? and no i really dont think so its good for training an rx 570, google colab t4 free daily gpu is more powerful
half cove
Hello Why I get that error, someone know?
error seems super weird
might be config error
low shard
Can someone help me with a JSONDecodeError
send google colab
willow bronze
error seems super weird
ohh okey
low shard
outdated colab thats why
alpine juniper
Ok
half cove
if not it might just be that the config for tts isnt installed
alpine juniper
Could you send me a better one?
low shard
Could you send me a better one?
use the updated ones you can find in the cloud folder inside the rvc folder of our docs https://docs.aihub.wtf
read which is the best for your needs, every of them has pros and cons written
alpine juniper
Thank you
low shard
yw
elder estuary
yea we arent affiliated with kits.ai bc its greedy and focus more on money than ...
i cant even find my kits.ai model anymore
rare gobletBOT
Ayo? @elder estuary level 2 !!!
low shard
i cant even find my kits.ai model anymore
kits.ai models arent allowed here anymore, we arent affiliated with them
elder estuary
kits.ai models arent allowed here anymore, we arent affiliated with them
no on kits.ai i made a model, cant find it anymore
on kits
junior flame
What should I do if when I talk to someone with the voice changer turned on, there are some voices behind me, like a trail?
thin stump
im areactive what? and no i really dont think so its good for training an rx 570...
I thought you knew a lot about applio so I just told him to post here, my bad
proven hill
someone needs help?
simple igloo
can anyone help me setup ASIO Audio rerouting or whatever its called?
proven hill
this new layout is so confusing
wich one?
simple igloo
can anyone help me setup ASIO Audio rerouting or whatever its called?
downloaded the installer from github but I dont understand how to use it in RVC
proven hill
the RVC one
oh look ilaria rvc, i wonder who ilaria is, you should probably ask her
fiery raptor
wheres the conversion button to convert audio files 😭
fiery raptor
oh look ilaria rvc, i wonder who ilaria is, you should probably ask her
Hi ilaria 
wispy escarp
I need help too pls, i'm trying to use bridgit mendler's voice but the link doesn't work (I'm kinda new to this too)
fiery raptor
Hi ilaria <:e_boorgir:1159654275069255750>
is it this button?
proven hill
Hi ilaria <:e_boorgir:1159654275069255750>
fiery raptor
OHHHHHHH
proven hill
its a giant button i dont know how did you miss it AUHUAHSAHUSASHU
fiery raptor
i use a red light thingy