clicked X, refreshed model, selected model (pth) , dragged audio file, clicked convert
still error
#✨│ai-help
worldly finch
half cove
Have you tried disconnecting and deleting runtime and after putting audio don't click refresh on audio
worldly finch
start execute is still running it is not disconnected
ye i dont click on refresh audio
do i need to disconnect everytime? when i want to convert new audio
half cove
Not on pc rn so I can't test ila rvc
half cove
do i need to disconnect everytime? when i want to convert new audio
It shouldn't be like that
steel plover
start preprocess
['trainset_preprocess_pipeline_print.py', 'C:\Users\Kaiser\Desktop\RoyceAI\vocalsv23.7.0\datasets\', '32000', '11', 'C:\Users\Kaiser\Desktop\TRAIN\Mangio-RVC-v23.7.0/logs/RoyceAI', 'False']
Fail. Traceback (most recent call last):
File "C:\Users\Kaiser\Desktop\TRAIN\Mangio-RVC-v23.7.0\trainset_preprocess_pipeline_print.py", line 118, in pipeline_mp_inp_dir
for idx, name in enumerate(sorted(list(os.listdir(inp_root))))
FileNotFoundError: [WinError 3] The system cannot find the path specified: 'C:\Users\Kaiser\Desktop\RoyceAI\vocalsv23.7.0\datasets\'
end preprocess
i remplaced the file names to the ones that the error was giving and so far dosent really seem to change . now iam lost .3.
@half cove toughts ?
proper shale
yeah put the dataset folder inside mangio
steel plover
gotcha ima try that
worldly finch
It shouldn't be like that
what do i need to do for the next time? if i want to convert a new audio after i already did one
steel plover
C:\Users\Kaiser\Desktop\TRAIN\Mangio-RVC-v23.7.0\datasets\RoyceAI\vocals23.7.0\datasets
it kidna looks like this , is good ?
half cove
what do i need to do for the next time? if i want to convert a new audio after i...
There should be a x on the audio uploaded you can click that then put new audio
azure marshBOT
RVC Guides (How to Make AI Cover)
half cove
C:\Users\Kaiser\Desktop\TRAIN\Mangio-RVC-v23.7.0\datasets\RoyceAI\vocals23.7.0\d...
Try not make the file length too long
steel plover
works :v
worldly finch
There should be a x on the audio uploaded you can click that then put new audio
yea when i click x on previous audio and than drag a new one
i get error
half cove
Weird
steel plover
Try not make the file length too long
ty , ill keep that in mind :3
half cove
It might have problems rn
molten pecan
i get error
Try using another model
half cove
ty , ill keep that in mind :3
wispy canyon
Why does every google drive model direct me to a page that says the file doesnt exist or incorrect url?
rare gobletBOT
Ayo? @worldly finch level 11 !!! 
molten pecan
I see
rare gobletBOT
Ayo? @wispy canyon level 1 !!!
half cove
Does it convert at all or does it only break after putting new audio
half cove
Ohh
wispy canyon
sends me to this
worldly finch
Does it convert at all or does it only break after putting new audio
after putting new audio
half cove
The original user may have deleted it
half cove
after putting new audio
I'll test later if I have time ig
wispy canyon
Hmm fair enough. Just seems wierd its the same story with EVERY google drive one
worldly finch
it worked after i stopped the runtime
half cove
Hmm fair enough. Just seems wierd its the same story with EVERY google drive one
Most Google drive ones are old
wispy canyon
oh i gotcha. Thank you
worldly finch
now i want to convert another audio so i click X on audio
and drag the new one
see
worldly finch
now i want to convert another audio so i click X on audio
poopmasta
half cove
It's prob ila rvc having problems
half cove
As I said I am on mobile right no so I can't check
worldly finch
it worked after i stopped the runtime
.
worldly finch
now i want to convert another audio so i click X on audio
.
proper shale
i honestly don't know then
worldly finch
it only 1 audio each runtime of execute
steel plover
umi forgot o install tensorboard.. anyone can tell me how to do it ina nutshell pls ? :3
worldly finch
https://colab.research.google.com/github/iahispano/applio/blob/master/assets/App...
kk i will try
steel plover
the page for ROD is down so.. idk where to go
rare gobletBOT
Ayo? @steel plover level 2 !!!
proper shale
umi forgot o install tensorboard.. anyone can tell me how to do it ina nutshell ...
put this bat inside ur folder, run it to install tb and run
ur rvc folder*
steel plover
put this bat inside ur folder, run it to install tb and run
ty !
proper shale
in case shit breaks, delete tensor_venv folder and re run the bat
steel plover
in case shit breaks, delete tensor_venv folder and re run the bat
i guessi ahve to stop the trianing i have ongoing too right ?
proper shale
i guessi ahve to stop the trianing i have ongoing too right ?
nope
proper shale
yw
ornate hatch
It's taking a looot. Is that normal?
hot lagoon
We all good 
proper shale
It's taking a looot. Is that normal?
if you're on huggingface maybe
ornate hatch
if you're on huggingface maybe
How much does it take normally on huggingface?
proper shale
seems to work , do i have to use the same bat FOR WHEN IW ANA USE tb ?
yeah
proper shale
How much does it take normally on huggingface?
takes a while depending on ur audio's length, just wait
ornate hatch
Oh, okay ty. Thanks for helping
proper shale
You're welcome :)
hot lagoon
We all good <:nails:1159569314848972891>
proper shale
yeah ur doing good
steel plover
uhh..
for some reasonmy steps went up to 5k
i think i have done something wrong :v
proper shale
nah ur graph doesn't seem that bad
steel plover
nah ur graph doesn't seem that bad
i stoped it at 120 steps and sounded like . really ass
this is the voice
proper shale
i stoped it at 120 steps and sounded like . really ass
yeah ofc, 120 steps is meh
do u mean epochs
steel plover
yeah ofc, 120 steps is meh
so i just should let it run >
steel plover
do u mean epochs
yeash epocs.
proper shale
yeash epocs.
oh that makes more sense
steel plover
so i just let it run ?
proper shale
so i just let it run ?
yeh, keep going and check the model once u think you're overtraining
steel plover
i set it to 350 epocs first .
steel plover
yeh, keep going and check the model once u think you're overtraining
gotcha but . idk if this is undertraining or overtraining
proper shale
gotcha but . idk if this is undertraining or overtraining
currently not overtrained
steel plover
pitch the audio down using the RVC gui, i think it should sound beter
seems like it continued rather than starting over..is it good o r irather start over?
young halo
Yo, I'm using tensorboard with my training, I followed every single step but it's no showing anything at all
half cove
Yo, I'm using tensorboard with my training, I followed every single step but it'...
If you trained / are traning a model it should show up after clicking refresh
hot lagoon
We all good <:nails:1159569314848972891>
Dose it look find or nah? 
young halo
If you trained / are traning a model it should show up after clicking refresh
I did so, and I set up that automatically also but it's the same
proper shale
Dose it look find or nah? <:joe_what:1169303301163077672>
as i said, yes
rare gobletBOT
Ayo? @young halo level 7 !!!
half cove
I did so, and I set up that automatically also but it's the same
I dont see anything wrong try asking in #1192011222023950368 so more people see it
proper shale
I did so, and I set up that automatically also but it's the same
zoom out in loss g total
young halo
I dont see anything wrong try asking in <#1192011222023950368> so more people se...
Is that orange line that indicates overtraining points, it's not showing
proper shale
and see if theres anything
young halo
Oooooo
proper shale
zoom out in loss g total
alt + zoom once you're over it
half cove
Was that it 😭
young halo
alt + zoom once you're over it
That actually helped me, tysm!
proper shale
That actually helped me, tysm!
yw! ;)
half cove
young halo
Ok, I input 2000 Epochs for this model as I read on a guide, and the OT point is at around 1k Epochs, do you recommend me to stop it at around 300 Epoch or until it's near the OT point?
proper shale
Ok, I input 2000 Epochs for this model as I read on a guide, and the OT point is...
If you're on Colab you gotta stop and train on another session
After like 1 or 2 hours
But yeah, train until you feel it's been overtraining for a while
Then you can check the models
steel plover
thisis going crazy
its going up the loss
@proper shale
i dont think the graph at the bottom should be going up .3.
molten pecan
i dont think the graph at the bottom should be going up .3.
Yeah seems to be time to stop
steel plover
ok so what epoc wopuld be this one ? .3.
molten pecan
ok so what epoc wopuld be this one ? .3.
Like I said in the guide, the lowest point
Zoom in and see which one it is
molten pecan
or what am i looking for exactly ? <@791013455251374151>
Is that loss/g/total?
steel plover
Is that loss/g/total?
yeah
rare gobletBOT
Ayo? @steel plover level 3 !!!
steel plover
steel plover
If that's the lowest point, remember the Step number
what is exactly the step number ? , whats kinda what iam missing .3.
since so far is on a diferent setting i think .
molten pecan
what is exactly the step number ? , whats kinda what iam missing .3.
MJ how do you explain this 
proper shale
what is exactly the step number ? , whats kinda what iam missing .3.
9.214 is the step number
or whatever was cited earlier
also, try out the latest version of the model too
molten pecan
You can see the Step in the black window on the left
steel plover
in betwen these i supose ?
molten pecan
The window here
steel plover
also, try out the latest version of the model too
what would be the latests? . ij ust downloaded and foillowed the latest vid on how to download the rvc
mango
proper shale
what would be the latests? . ij ust downloaded and foillowed the latest vid on h...
the latest file in weights folder
steel plover
the latest file in weights folder
oho k gotcha
steel plover
But they overtrained already
yeah tahts what i mean xd
proper shale
But they overtrained already
how do you know
molten pecan
how do you know
Scroll up
proper shale
g graph going up is not necessarily overtraining lmao
molten pecan
They showed a full pic of TB
molten pecan
g graph going up is not necessarily overtraining lmao
It's been going up for a while already
proper shale
It's been going up for a while already
either way, test out the latest version
it can (and most likely will) sound better
molten pecan
That's fair
steel plover
test latests ig .
thin sparrow
Hello. Question, does the applio vthree include the pretrain type? The Ov2 or the Rin_E3?
half cove
Hello. Question, does the applio vthree include the pretrain type? The Ov2 or th...
You gotta add em in
It doesn't come preinstalled
thin sparrow
and I was also referring to the collab
thin sparrow
ok
proper shale
Hello. Question, does the applio vthree include the pretrain type? The Ov2 or th...
nope, not on colab
worldly finch
i need helpp
proper shale
test latests ig .
again, pitching down could make it sound better
steel plover
again, pitching down could make it sound better
it already is , -8
worldly finch
everytime when i want to convert a another audio i get error
proper shale
it already is , -8
try -12
worldly finch
i can convert 1 audio per execute
thin sparrow
Somthing like this? <https://colab.research.google.com/github/poiqazwsx/RVC-Note...
this still working?
steel plover
-12 pitched down. still soudns too robotic. .3.
e 315
@proper shale
iam lost.
pls
@half cove ? :3
half cove
Can't listen to audio cuz on mobile rn xd
steel plover
well is too robotic at the end
should sound more like this https://www.youtube.com/watch?v=0DviyuGNeWc&t=373s
proper shale
-12 pitched down. still soudns too robotic. .3.
its probably the audio you're using
doesnt sound like the character
steel plover
its probably the audio you're using
from mp3 to FLC
steel plover
in terms of pitch
kinda varies but it is -12 pitch so it sounds more deep than how iti s
but opriginaly sounds more like this
pitch : 0
molten pecan
but opriginaly sounds more like this
Listen, how about you test it with the lowest training point?
Instead of the most recent
steel plover
Listen, how about you test it with the lowest training point?
yeah mima use that instead
molten pecan
@proper shale watch me being right >:3
proper shale
<@622856015444049937> watch me being right >:3
objection: will sound the same because it's not the model, it's the audio they're using to convert
steel plover
proper shale
molten pecan
The closest to before the OT point
proper shale
then use e150
steel plover
this is the audio . i used for the training .
molten pecan
Only 10 secs? That's a lil short
proper shale
bruh what
steel plover
Only 10 secs? That's a lil short
is 15 mins long.
molten pecan
Ah
proper shale
@molten pecan r u smoking
proper shale
again, the audio they're using to convert is the problem
steel plover
time to cross out fingers.
steel plover
again, the audio they're using to convert is the problem
then how do i fix it ?
proper shale
again, the audio they're using to convert is the problem
They should use another audio that's more whispery
steel plover
so far on the vid told me FALC
proper shale
so far on the vid told me FALC
first off, nuh uh, it's still an mp3 because you downloaded it from YouTube, you're not making audio quality better
second, the Audio you're using to test is the problem
steel plover
first off, nuh uh, it's still an mp3 because you downloaded it from YouTube, you...
ok , then how do i do so its high quality then ?
proper shale
it's not whispery, not similar to the speaker
steel plover
it's not whispery, not similar to the speaker
true .ima make anew audio for it .
proper shale
it's not whispery, not similar to the speaker
and since the voice model is monotonous, its just gonna sound bad
proper shale
true .ima make anew audio for it .
yeah that should make it better
steel plover
rare gobletBOT
Ayo? @steel plover level 4 !!!
steel plover
a tat better but still feels a bit robbot like
worldly finch
i have a error too
proper shale
pitch that down and i think it might sound good ngl
worldly finch
maybe someone knowns
each execute run i can do 1 convert
after 1 convert is done i click to X on audio
and drag another audio to convert
steel plover
thisis darkmagic
proper shale
thisis darkmagic
do u mean this in a good way or
steel plover
do u mean this in a good way or
yeah, but do i always have to keep it pitched down liek that and talk in the same way ?
proper shale
ok, give us the ila rvc cell output once u do conversion
steel plover
i mean iwant ti obe flexible so it can work with ceratain acentuations .
proper shale
yeah, but do i always have to keep it pitched down liek that and talk in the sam...
i mean u gotta pitch down yeah
and sorta be monotonous
proper shale
i mean iwant ti obe flexible so it can work with ceratain acentuations .
yeah uh, it ain't gonna be flexible, from what I've heard the character is very monotonous
steel plover
true .
but thats is kinda what iam going for regadless
i guess i can call this a succes mostly due to the robot voice part .
but about the MP3 thingy , how do i geta higher quality for the ytb ? @proper shale
worldly finch
ok, give us the ila rvc cell output once u do conversion
proper shale
but about the MP3 thingy , how do i geta higher quality for the ytb ? <@62285601...
you can't
thats why u shouldn't really go for yt vids, maybe as a last resort
steel plover
i guessi cant get better from this ?
steel plover
thats why u shouldn't really go for yt vids, maybe as a last resort
then from ?
proper shale
just feels like a issue with ila rvc, try applio
https://colab.research.google.com/github/iahispano/applio/blob/master/assets/Applio.ipynb
proper shale
then from ?
depends on the medium
steel plover
well there is only this ytb vid with the clear audio .
proper shale
depends on the medium
search and you'll find audios from games n shit
steel plover
ima try and use stacher
steel plover
I just use stacher
what format in stacher?
proper shale
@half cove you do realize that this doesn't really help right?
proper shale
it's like polishing a jpeg
steel plover
U gotta install ffmpeg first
seems like it didnt required me to iunstal ffmpeg
proper shale
yt audio works for training models, but the audio quality is limited regardless of file type
half cove
yt audio works for training models, but the audio quality is limited regardless ...
At like 40k
proper shale
At like 40k
Oh and, 192kbps opus audio
half cove
covert turret
I just checked Tensorboard and apparently the step 12.6k is the best step before my model overtrains, my question is: how do I figure out what epoch number that step is?
for reference, step 12.35k is 650 epochs in my model
proper shale
I just checked Tensorboard and apparently the step 12.6k is the best step before...
check the weights folder
and see what's closest
also, u should try the latest model too
covert turret
and see what's closest
650 is what I saw in the folder
steel plover
this is considereda felony at this point @half cove
half cove
On your pth every checkpoint has epoch then steps apart from last one
proper shale
650 is what I saw in the folder
yeah then that's it
covert turret
but i guess it needs a bit more to get there.
half cove
this is considereda felony at this point <@936673139419664414>
Listen to mj
proper shale
this is considereda felony at this point <@936673139419664414>
Why
proper shale
384khz for a YT audio???
steel plover
tim to see wherei can geta clearler audio
covert turret
yeah then that's it
650 is 12.35k steps, so ig i should do 660 or something?
proper shale
650 is 12.35k steps, so ig i should do 660 or something?
yeah either that or 650
covert turret
yeah either that or 650
gotcha, thank you!
proper shale
yw :)
steel plover
well then i guess it is settled , but what can i do now to improve what i have ? re-train and see if iget something better ?
or am i just stuck with it ?
proper shale
well then i guess it is settled , but what can i do now to improve what i have...
nah the model is decent
there is nothing much u can do ig
steel plover
yeah.
what about other languages?
for example if i speak in spanish and wish t the voice to be use spanish do ihave to do something else o ?
worldly finch
just feels like a issue with ila rvc, try applio
https://colab.research.google....
how does this works i m unkowns with this
proper shale
nah u dont need to do anything else
it will work out just fine, might have a weird accent tho
worldly finch
never used before
proper shale
how does this works i m unkowns with this
its straightforward, just run the steps and see
worldly finch
insert link there, and click descargar modelo
i got error before re did again now worked
proper shale
woo
now just refresh the model thingy in inference section, upload ur audio and u should be good
worldly finch
yes it worked but i wonder something
since start
the output voice isnt same as my voice
proper shale
i mean... its a voice conversion
worldly finch
yea
proper shale
ofc it wouldn't sound like ur voice...?
proper shale
unless u trained urself
proper shale
^
again, you dont need to do anything extra
answered this b4
it might sound weird tho
worldly finch
so own sound output will not give my exact same voice
proper shale
so own sound output will not give my exact same voice
i mean, yeah
its converting ur voice in some way
steel plover
it might sound weird tho
ho so i just have to talk in spanish and will talk in like spanish accent ? :0
worldly finch
i mean, yeah
its converting ur voice in some way
alr now everything is fine
thankss
proper shale
i see now ty
and yw!
like, i believe it will sound like an american trying to speak english
u can disable index tho, it might sound better that way
steel plover
u can disable index tho, it might sound better that way
how i do that ?
proper shale
how i do that ?
just set the index slider to 0
steel plover
just set the index slider to 0
on the training or on rvc ?
proper shale
on the training or on rvc ?
on rvc yea
half cove
Is that rvc gui?
molten pecan
thats what i been using for it .
Oh nooo
molten pecan
It's too outdated
steel plover
let me gues
molten pecan
half cove
Its only and doesn't have things like rmvpe
molten pecan
HAUSJDHASKFNSA
https://aihubdocs.github.io/en/ use mainline rvc
molten pecan
Huh? <:nails:1159569314848972891>
The keysmash
molten pecan
@steel plover be more careful with the keysmash next time 😭
steel plover
ty
yeha so i was using something old and now i gota re-learn how ot use the rvc new one .
great :v
proper shale
@steel plover pls use mangio for inferring to
too*
its pretty good
just gotta put ur file paths n stuff for the audio
steel plover
<@209046076773105678> pls use mangio for inferring to
iam using magio for inferring
proper shale
yea this is the one
proper shale
and i guess this si the right one ?
this is original rvc
steel plover
this is original rvc
so , good , right?
steel plover
bet.
rare gobletBOT
Ayo? @steel plover level 5 !!!
steel plover
my whiole discord layout just changed
half cove
Huh?
steel plover
idk some chanels got moved or removed .
the SD one seems to onl;y have "help-ai-art and nothing else
steel plover
this is taking a hell ofa while
half cove
I mean you could just use mangio
steel plover
I mean you could just use mangio
i dont wana mix up the training one with the use one :v
half cove
Do you have the infer only mangio?
steel plover
ok where od i put my trained models ? , and why is everything on spanish ? .3. , i rahterh ave it on english tbh
and some text is on asian
iam so confused.
half cove
ok where od i put my trained models ? , and why is everything on spanish ? .3. ,...
Put pth in assets then weights and index in logs
half cove
and some text is on asian
First is infer second is train
steel plover
First is infer second is train
well but iwana make my whole RVC be english .3.
half cove
well but iwana make my whole RVC be english .3.
You could use mangio train and infer
Version
steel plover
You could use mangio train and infer
infer vers is also in spanish'
but icna manage to understand and co-relate
half cove
Spanish?
half cove
Where did you install it?
half cove
my c drive .3.?
Link xd
pine hill
this chat is for the colab version?
half cove
-links
azure marshBOT
half cove
this chat is for the colab version?
What colab you need
half cove
What colab you need
Voice changer?
Making covers?
Training?
half cove
i dont understand .3.
Like the place you downloaded it from
pine hill
Voice changer?
Making covers?
Training?
holly, i just wanna try the voice changer
young halo
I know I make a lot of questions lol, but I just want to know, if I need to send my models to QA before posting them even though I have the model maker role
half cove
holly, i just wanna try the voice changer
half cove
is from theyr OG links.
It should be in English so I'm confused lol
pine hill
<https://colab.research.google.com/github/hinabl/voice-changer-colab/blob/master...
i need to download something or just with the web?
steel plover
It should be in English so I'm confused lol
ikr?
isthere way i can change the language regadless?
half cove
i need to download something or just with the web?
To use it in a game you need a virtual audio cable
pine hill
To use it in a game you need a virtual audio cable
i just wanna use in discord with friends
half cove
isthere way i can change the language regadless?
Could just try translate page ig
steel plover
Could just try translate page ig
ikd howo to do that on opera gx xd
young halo
Hm?
I mean, my model got approved yesterday when I sent it to #1159514067187277865 and I got the model maker role, so I was wondering if I need to submit my models in there before posting them in #1175430844685484042
pine hill
To use it in a game you need a virtual audio cable
but i don't understand, i just run it or i need to download something?
rare gobletBOT
Ayo? @pine hill level 1 !!!
molten pecan
but i don't understand, i just run it or i need to download something?
Mejor pregunta en #🌍│español
Eddy podrá ayudarte
half cove
Might long to respond discord bejgn a bitch rn
steel plover
Mejor pregunta en <#1159369117854347276>
this is wherei puit the path tothe audio i wana infer right ? .
pine hill
Mejor pregunta en <#1159369117854347276>
soy medio bobi, no vi ese canal, gracias
half cove
I mean, my model got approved yesterday when I sent it to <#1159514067187277865>...
After you post the model it's checked by staff
molten pecan
this is wherei puit the path tothe audio i wana infer right ? .
Si. Porfa lee la guía de uso
young halo
After you post the model it's checked by staff
Oooo I got it, if there's any problem with it, does it get taken down or something?
molten pecan
Oooo I got it, if there's any problem with it, does it get taken down or somethi...
Yes. Check the model maker page
You'll find the reasons that can get your model rejected
young halo
Yes. Check the model maker page
Tysm!
molten pecan
yw!
half cove
Ok I'm back
half cove
but i don't understand, i just run it or i need to download something?
This should walk you through it
half cove
You'll find the reasons that can get your model rejected
He's asking when posting a model as model maker lol
brittle wing
how me make ai voice do funny stuff
stark abyss
i will translate
half cove
how me make ai voice do funny stuff
Are you trying to use voice changer or making ai covers
half cove
i will translate
?
brittle wing
Are you trying to use voice changer or making ai covers
ai covera
stark abyss
how me make ai voice do funny stuff
translation: M̷͐̒Ë̵́̓ ̴̛͍N̵̰̏E̸̛̦E̵͐̓D̵͒ PRO̷͍͝G̷͗̀RAM̶̛͛ ̸̛́N̴̓͗O̸W GIVE̶͒͆ ̸ME I N̵͖͗E̴̔ED DA F̷͊̚Ȗ̵̕N̷̞͑Ṋ̸̓Ÿ̸̹́
molten pecan
Use the search bar to find the part about making ai covers
It's on the landing page
No need to do dat
half cove
It's on the landing page
Ooo nice
brittle wing
me no know how it work no make sense UISDHGBEAVBFVHSAXCBSDFJGSDGBSDG
rare gobletBOT
Ayo? @brittle wing level 1 !!!
molten pecan
me no know how it work no make sense UISDHGBEAVBFVHSAXCBSDFJGSDGBSDG
Read it. I made it beginner-friendly
brittle wing
Read it. I made it beginner-friendly
no
half cove
half cove
Ok so your using something called rvc1006 something right?
half cove
Go on your rvc folder
sinful lava
okkk
half cove
idk what or where the pth files are
Did you download a model that you want to use?
half cove
Well first are you trying to make a ai cover or a ai model?
sinful lava
i did my voice and filterd it
sinful lava
ima make ai model then make me sing a song lmao
sinful lava
i have the program ope n
half cove
You can either make one or get one from #1175430844685484042 or weights.gg
sinful lava
naw make one
but like
Im stuck on what the pth file is
maybe i missed something
i listend to all steps
half cove
I think your on the part about using an ai model
half cove
Yep
sinful lava
so i do that thank i should do the pth and stuff
half cove
Scroll down and it should have the tutorial for training and stuff
half cove
so i do that thank i should do the pth and stuff
After doing it you should get index and pth
sinful lava
alright ill keep you updated poop master
half cove
Just use the inference part of the tutorial after you make it
molten pecan
Just use the inference part of the tutorial after you make it
Woah the site is pretty useful
And to think it's early in development
half cove
sinful lava
Just use the inference part of the tutorial after you make it
where is this 🥲
rare gobletBOT
Ayo? @sinful lava level 2 !!!
half cove
where is this 🥲
The beginning is about inference the second bit is about training
It says in the titles
half cove
but icant find the url
You have to upload the model you created first or use a premade model
sinful lava
how do i upload my model
sinful lava
im sorry i dont wanna make you annoyed lolll
half cove
Dw
molten pecan
its gonna be the last chance to save them
😭😭😭😭
dawn nebula
im talking using my school laptop, connected to an amazon remote desktop service (1GB ram+1 core)
its so slow
switching servers take like 15 seconds
molten pecan
im talking using my school laptop, connected to an amazon remote desktop service...
Use discord canary
dawn nebula
whats that
azure marshBOT
RVC Guides (How to Make AI Cover)
molten pecan
is it web or local
Local
buoyant lotus
molten pecan
Pic goes hard but not related to help-rvc
Wait until you get perms and then post em on #✦│chat
buoyant lotus
aight boss
rare gobletBOT
Ayo? @buoyant lotus level 2 !!!
buoyant lotus
spare shale
mystic pawn
Is taking a little longer than expected
Maybe it's a heavy model
Any suggestions why?
brittle wing
anyone doing Google Collab & RVC?
i need help, this hurts my head
im getting
[Errno 2] No such file or directory: '/content/voice-changer/server'
/content
ModuleNotFoundError Traceback (most recent call last)
<ipython-input-2-e8b7f718dd62> in <cell line: 24>()
22 get_ipython().run_line_magic('cd', '/content/voice-changer/server')
when doing Start Server using ngrok
molten pecan
does somebody know how much powerful gpu do i need
I heard the RTX 3060 series onwards was the best option
But you can execute RVC through the cloud though
plush heron
does anyone know how to get the old rvc thing where it was its own like app thing with all the options instead of it being on a browser? I factory reset my computer recently and dont remember how to get it back
where like you had to download it from drive
mystic pawn
Well, things had changed a little
Basically we have Rvcdisconnected you train there and you use another Inference cloud space
molten pecan
does anyone know how to get the old rvc thing where it was its own like app thin...
RVC GUI? The vertical blue window?
plush heron
RVC GUI? The vertical blue window?
yeah
rare gobletBOT
Ayo? @plush heron level 1 !!!
plush heron
it was easy to use tho
molten pecan
it was easy to use tho
Yea but it gave lower quality results, very few settings, and had only obsolete models
I have a usage guide of the modern RVC for newbies
https://aihubdocs.github.io/en/ check the mainline page
;3
plush heron
alr thanks
molten pecan
alr thanks
Though if you want it even more simple check the Ilaria RVC one
Have fun!
modern surge
-colab
vast forge
How do I make a model?
wispy granite
yo can i run RVC-GUI.bat is it good and safe me bad at illiterate
foggy pumice
On mobile and everything else is fine but I can’t upload an audio
vast forge
yo can i run RVC-GUI.bat is it good and safe me bad at illiterate
yes thats what you open
rare gobletBOT
Ayo? @vast forge level 1 !!!
vast forge
wispy granite
thx
vast forge
you just wait for it to open it could take a few minutes
im working on making my first model rn
wispy granite
thank you
vast forge
your welcome
spiral osprey
When you make a voice model, you make a dataset with several songs with a clean voice, right? but I can make a dataset with the same song over and over until it reaches 40 minutes. For example, I have a song and I decided to press ctrl c and v until I reach a certain amount of music that lasts 40 minutes. Is it possible to do this?
molten pecan
When you make a voice model, you make a dataset with several songs with a clean ...
You can do it, but it won't give any improvement
Variation in tone is very important for a great model
spiral osprey
You can do it, but it won't give any improvement
so it's good to have several songs to make a good voice model.
molten pecan
so it's good to have several songs to make a good voice model.
Yeah. Check the datasets page for more tips
=3
spiral osprey
https://aihubdocs.github.io/en/
I've seen that.
royal jacinth
does anyone know a Google collab that uses ov2super pretrain
real igloo
uhhhhh
so after the error i described my gpu utilization is now at 4.2billion
it's legit like double the 32 bit integer
💀
also how do i fix this i can't do feature extraction
azure marshBOT
RVC Guides (How to Make AI Cover)
dawn nebula
does anyone know a Google collab that uses ov2super pretrain
-colab
dawn nebula
rvc disconnected
terse cairn
I need train new model how can I doit
dawn nebula
also how do i fix this i can't do feature extraction
send error log
terse cairn
What that means
dawn nebula
I need train new model how can I doit
good luck reading this, the tutorial is straightforward
dawn nebula
What that means
sorry i was helping another person
terse cairn
I have index file and model path but I don't what to do next so can you tell me what to do nex
dawn nebula
I have index file and model path but I don't what to do next so can you tell me ...
it is in the website, please search
we helpers are not meant to spoonfeed
terse cairn
Ok sorry and thanks
rare gobletBOT
Ayo? @terse cairn level 1 !!!
brittle wing
Someone ruined AI Cover generator web UI the outputs sound so God awful
frigid vapor
im reading the rvc guide and it says i need rtx nvidia
i have rtx radeon is that fine?
sinful lava
im making my own voice
and i have everything recorded
but when it says upload voice model with the .pth im not sure how to make that
any help?
dawn nebula
you need to train it first
https://aihubdocs.github.io/en
check the training parts
you already have a dataset, now you only need to train
sinful lava
how train doe
latent kettle
how train doe
You need to train your model
sinful lava
ik but where do i do that
sinful lava
You need to train your model
?
latent kettle
ik but where do i do that
Rvcdisconnected, by kit lemonfoot
latent kettle
and make it like the pth file and index
Yep
sinful lava
okkk thanks
latent kettle
Just record your voice make a zip and upload it to Google drive
sinful lava
got it
azure marshBOT
Suggestions for @sinful lava
dense quartz
hey guys, is Applio application for tts? Ever tutorial I found on youtube i cant follow entirely bsc the step is very differently
forest sluice
does anyone know a Google collab that uses ov2super pretrain
rvc disconnected
copper pebble
hey guys, is Applio application for tts? Ever tutorial I found on youtube i cant...
use 11labs or ilaria rvc for tts
dense quartz
use 11labs or ilaria rvc for tts
But 11labs can’t use voice model right? What if I wanted something that can use voice model
copper pebble
actually theres a method to use voice model directly in the tts
just download the speech and convert it on RVC
dense quartz
just download the speech and convert it on RVC
Oh ok, sorry I might be new to this field, what should I be using to download the speech and convert it?
Bsc through the voice model in this discord server I realise bsc can use their own created voice model to become a tts
dense quartz
Bsc through the voice model in this discord server I realise bsc can use their o...
So was wondering how did they did it
copper pebble
Oh ok, sorry I might be new to this field, what should I be using to download th...
create ur speech on any TTS
like 11labs, ilariatts etc
download the speech generated
then put it to RVC
put ur model and convert it
u can use ilaria rvc also for the convertion
glad zealot
like 11labs, ilariatts etc
ilariarvc can do rvc model tts voice
copper pebble
ilariarvc can do rvc model tts voice
ilaria tts is integrated in ilaria rvc
copper pebble
u can use ilaria rvc also for the convertion
like i said
glad zealot
yes
dense quartz
i see this is a intersting idea! my current situation is i have no idea how do i put the voice model inside the illaria RVC
dense quartz
u can use ilaria rvc also for the convertion
i can use a lot of people voice, but how do i import the voice model?
copper pebble
i can use a lot of people voice, but how do i import the voice model?
u can find guides on ilaria rvc here #1159513888199540817
to import tho u just need the download link of the model > go to "download model" > paste the link & download it
dense quartz
u can find guides on ilaria rvc here <#1159513888199540817>
ah thanks a lot! I will go through it
glad zealot
from my personal testing, mainline seems to train faster than applio, tho its only by a few seconds every epoch
not sure about the results on others tho
azure marshBOT
RVC Guides (How to Make AI Cover)
oblique venture
Any way to run realtime tts with rvc model? i want to put a voice for my chatbot
sand iris
Any way to run realtime tts with rvc model? i want to put a voice for my chatbot
the voice-changer is able to get audio input from desktop audio
But you'll need to have your tts playback on a dummy sound output so that you don't hear both source and converted audio
That or use virtual audio cable the opposite way we usually do. (TTS output to virtual audio cable, voice changer picks up audio from virtual audio cable and plays back to desktop)
forest sluice
That or use virtual audio cable the opposite way we usually do. (TTS output to v...
Anyway, how do you get the voice changer on Linux? jack2?
sand iris
Anyway, how do you get the voice changer on Linux? jack2?
In terms of audio ?
It unless requires on the browser, so whatever that supports
I'm not sure how it does it when using "server". I think it's portaudio, but I haven't looked at what that supports
PortAudio is a free, cross-platform, open-source, audio I/O library. It lets you write simple audio programs in 'C' or C++ that will compile and run on many platforms including Windows, Macintosh OS X, and Unix (OSS/ALSA).
oblique venture
the voice-changer is able to get audio input from desktop audio
Can it run with an acceptable performance on cpu?
sand iris
Can it run with an acceptable performance on cpu?
No
oblique venture
fierce isle
if i need to separate 2 people singing together what do i use on UVR5
molten pecan
if i need to separate 2 people singing together what do i use on UVR5
UVR-BVE
brave garnetBOT
⠀
Google Colabs 
⠀
AICoverGen-WebUI
Useful for making quick covers, by Hina.
AICoverGen-NoWebUI
Useful for making covers, doesn't inclued a UI, by Ardha, by Eddy, Hina and Gdr.
RVC Disconnected
To train new voice models, by Kit Lemonfoot.
EasyGUI
The OG interface, by Rejects.
⠀
Made with love by Ilaria 💖
rare flint
i assume training model locally requires at least RTX card?
proper shale
i assume training model locally requires at least RTX card?
i mean, it's more of a recommended setting
training with 10 series is slow
rare flint
i have 16 series
rare gobletBOT
Ayo? @rare flint level 1 !!!
rare flint
but as long it works i assume it works yes?
i do remember training with colab took 4-6 hours
proper shale
i have 16 series
if it's the 1650 I would recommend going for Colab
idk about the 1660, probably slower than 20xx series
rare flint
damn
proper shale
i do remember training with colab took 4-6 hours
it really depends, you have to limit some of the sessions on the newer Colabs to prevent losing filws
brave garnetBOT
Made with love by Ilaria 💖
rare flint
how does huggingface one works?
radiant whale
Can somebody make a website similar to YouTube which free copyright strike or termination?
So we can post our ai cover
molten pecan
how does huggingface one works?
It's outdated. Using the Colab space is much more convenient
Check the Ilaria RVC page :3
rose geyser
hi team! i just want to ask how to use the voice Changer client demo as an output for example when you are talking someone in discord
molten pecan
hi team! i just want to ask how to use the voice Changer client demo as an outpu...
rose geyser
thanksssssssssss
rare gobletBOT
Ayo? @rose geyser level 1 !!!
rose geyser
sorry, is this the page to ask for the Voice Changer client demo or in okada?
molten pecan
sorry, is this the page to ask for the Voice Changer client demo or in okada?
#🔍│help-w-okada is for both Okada and Voice Changer
(they are the same thing though)
_<
rose geyser
<#1159290161683767298> is for both Okada and Voice Changer
ok thank youuu
so it's working now, and i have another question if you don't mind
but the one that I am talking to, they also hear their own voice and it is also changing their voice, is there a fix about this or is this natural? thanks! i dont know what to search
sorry and thank you
molten pecan
ok thank youuu
so it's working now, and i have another question if you don't min...
I don't use the voice changer so idk
Good luck
rose geyser
alright ty
proper shale
ok thank youuu
so it's working now, and i have another question if you don't min...
The input output settings should be like this in W-Okada:
Input - Your mic
Output - CABLE Input
And in Discord:
Input - CABLE Output
Output: Your headphones
Make sure it's like that
So you don't get that annoying thing of converting the voice of ur friend
rose geyser
thank youuuu! ❤️
river heath
What are the best settings for a studio acapella reverb?
proper shale
What are the best settings for a studio acapella reverb?
if you're mixing, this really depends
just see what sounds better ig, just make sure not to drown the whole thing in reverb
wispy lava
File "G:\Artificial_Intelilgence\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\runtime\lib\site-packages\gradio\routes.py", line 437, in run_predict
output = await app.get_blocks().process_api(
File "G:\Artificial_Intelilgence\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\runtime\lib\site-packages\gradio\blocks.py", line 1349, in process_api
data = self.postprocess_data(fn_index, result["prediction"], state)
File "G:\Artificial_Intelilgence\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\runtime\lib\site-packages\gradio\blocks.py", line 1283, in postprocess_data
prediction_value = block.postprocess(prediction_value)
File "G:\Artificial_Intelilgence\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\runtime\lib\site-packages\gradio\components.py", line 2586, in postprocess
file_path = self.audio_to_temp_file(
File "G:\Artificial_Intelilgence\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\runtime\lib\site-packages\gradio\components.py", line 360, in audio_to_temp_file
temp_dir = Path(dir) / self.hash_bytes(data.tobytes())
AttributeError: 'NoneType' object has no attribute 'tobytes'```@sand iris
sand iris
👍
brave garnetBOT
⠀
Google Colabs
⠀
AICoverGen-WebUI
Useful for making quick covers, by Hina.
AICoverGen-NoWebUI
Useful for making covers, doesn't inclued a UI, by Ardha, by Eddy, Hina and Gdr.
RVC Disconnected
To train new voice models, by Kit Lemonfoot.
EasyGUI
The OG interface, by Rejects.
⠀
Made with love by Ilaria 💖
sand iris
```Traceback (most recent call last):
File "G:\Artificial_Intelilgence\Mangio-...
what do you press to trigger that btw ?
wispy lava
what do you press to trigger that btw ?
convert?
sand iris
so like infer audio ?
wispy lava
yeah
wispy lava
but gotta find where this gets called from
i can send you the functions from the traceback if you want
sand iris
i can send you the functions from the traceback if you want
wait there's more to it that what is above ?
wispy lava
wait there's more to it that what is above ?
no, i meant i could send u the full code of the funcs in the traceback if u wanted to check them out
hexed barn
Is Google Collab working?
wispy lava
like it has something to do with saving the file in the temp directory
but i cant figure out why its happening
sand iris
I have them on hand
wispy lava
hmmm, i wonder whatll happen if i just like pass
like add an exception and let the code run
sand iris
It shouldn't end with no data like it does here
https://github.com/Mangio621/Mangio-RVC-Fork/blob/main/infer-web.py#L236
Faulty shit might be there
Not suer about the gradio version used by the runtime, but it errors because of the inputs of this function
https://github.com/gradio-app/gradio/blob/v3.32.0/gradio/components.py#L360
wispy lava
yeah i can see it
the var
cant i just manually set it?
what is the value supposed to be exactly?
wispy lava
yep, cant do shit about that :///
sand iris
Would be great if I could get the issue here
wispy lava
maybe im missing smth
wispy lava
Bro had 5 issues one after each other and solved all of them 💀 das crazy
sand iris
Wait do you have that displayed?
OSError: Model file not found: hubert_base.pt
Or anything higher up that could be relevant?
wispy lava
oh actually i just noticed something up the traceback
rare gobletBOT
Ayo? @wispy lava level 4 !!!
wispy lava
File "G:\Artificial_Intelilgence\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\infer-web.py", line 292, in vc_single
audio = load_audio(input_audio_path0, 16000, DoFormant, Quefrency, Timbre)
File "G:\Artificial_Intelilgence\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\my_utils.py", line 103, in load_audio
raise RuntimeError(f"Failed to load audio: {e}")
RuntimeError: Failed to load audio: ffmpeg error (see stderr output for detail)```💀
how the f did i not notice that 💀💀
Defaulting to user installation because normal site-packages is not writeable
Requirement already satisfied: ffmpy in c:\users\imajo\appdata\roaming\python\python311\site-packages (0.3.1)```man idk what to do anymore
OH ACTUALLY
mangio has its own python and pip library
so yeah lemme try w that
sand iris
If you have ffmpeg available via PATH (no matter how) you should be good
It's available from winget if I remember correctly. That's like the one way that puts it in PATH for sure
winget install ffmpeg
wispy lava
heres the new traceback
File "G:\Artificial_Intelilgence\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\infer-web.py", line 300, in vc_single
if_f0 = cpt.get("f0", 1)
NameError: name 'cpt' is not defined
Traceback (most recent call last):
File "G:\Artificial_Intelilgence\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\runtime\lib\site-packages\gradio\routes.py", line 437, in run_predict
output = await app.get_blocks().process_api(
File "G:\Artificial_Intelilgence\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\runtime\lib\site-packages\gradio\blocks.py", line 1349, in process_api
data = self.postprocess_data(fn_index, result["prediction"], state)
File "G:\Artificial_Intelilgence\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\runtime\lib\site-packages\gradio\blocks.py", line 1283, in postprocess_data
prediction_value = block.postprocess(prediction_value)
File "G:\Artificial_Intelilgence\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\runtime\lib\site-packages\gradio\components.py", line 2586, in postprocess
file_path = self.audio_to_temp_file(
File "G:\Artificial_Intelilgence\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\runtime\lib\site-packages\gradio\components.py", line 360, in audio_to_temp_file
temp_dir = Path(dir) / self.hash_bytes(data.tobytes())
AttributeError: 'NoneType' object has no attribute 'tobytes'```this guy said use wav files not mp3
lets try that
same error
sand iris
ok now what the fuck
global n_spk, tgt_sr, net_g, vc, cpt, version
if sid == "" or sid == []:
global hubert_model
if hubert_model is not None: # 考虑到轮询, 需要加个判断看是否 sid 是由有模型切换到无模型的
print("clean_empty_cache")
del net_g, n_spk, vc, hubert_model, tgt_sr # ,cpt
hubert_model = net_g = n_spk = vc = hubert_model = tgt_sr = None
if torch.cuda.is_available():
torch.cuda.empty_cache()
###楼下不这么折腾清理不干净
if_f0 = cpt.get("f0", 1)
lemme link to the og repo
IT DELETES IT
THEN USES IT
WHAT THE FUCK
oh it's commented out
It didn't load the model apparently
wait no that wouldn't make sense
sand iris
this guy said use wav files not mp3
do you see any setting called "speaker id" ?
sand iris
try setting that speaker id thing to 0
if you ever see it
It should affect a value to cpt
wispy lava
Traceback (most recent call last):
File "G:\Artificial_Intelilgence\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\infer-web.py", line 6, in <module>
import config
ModuleNotFoundError: No module named 'config'
bro what 💀
that line doesnt exist
like it literally doesnt exist, line 6 is empty 💀
sand iris
python moment
wispy lava
Traceback (most recent call last):
File "G:\Artificial_Intelilgence\Mangio-RVC-v23.7.0_INFER_TRAIN\Mangio-RVC-v23.7.0\infer-web.py", line 25, in <module>
from config import Config
ModuleNotFoundError: No module named 'config'
wispy lava
Bro changed his mind now 💀
lol
idfk
i havent touched anything 💀
i just reran it
sand iris
python is lovely
wispy lava
true 💀 but how tf is that even possible
so like why tf is it giving me that error now?
like it was find before??? and i havent added that now, it was there from before
azure marshBOT
RVC Guides (How to Make AI Cover)
wispy lava
💀 but for some reason it decided : NAHHH SIKE i aint gonna work anymore
Brooo
i just reverted all the changes to code
and now it works
i mean it doesnt, but not that import error
UGH no embed 💀
anyways, that line is still fucking there but yeah it decides to work
weak needle
Is it normal for res to spike incredible high while playing games ?
Example: playing Monster Hunter World increases the res from ~60 to 25-30k ?
(GPU: RTX 4070 Ti
CPU: AMD Ryzen 7 7800X3D
RAM: 32gb, if that matters too.
since I saw it being mentioned here I tried exporting the model to onnx, but that didn't seem to do anything)
proper shale
Is it normal for res to spike incredible high while playing games ?
Example: pla...
Intensive games do take a toll on performance, so yeah
There isn't a lot to be done other than... maybe limiting framerate and setting graphics to low
wispy lava
Worse than I remembered
lol
dawn nebula
```
global n_spk, tgt_sr, net_g, vc, cpt, version
if sid == "" or sid == []:...
you need translation?
sand iris
Nah dw
rare gobletBOT
Ayo? @sand iris level 67 !!!
sand iris
I had the translation table opened
dawn nebula
lol
sand iris
Could just look up the corresponding English string from them lol
weak needle
There isn't a lot to be done other than... maybe limiting framerate and setting ...
Restarting my PC seems to have fixed whatever issue this was, it's at buf=100, res= 80 now while playing, strange >.<
Still thank you!
rare gobletBOT
Ayo? @weak needle level 2 !!!
proper shale
Restarting my PC seems to have fixed whatever issue this was, it's at buf=100, r...
Oh damn, that was quite weird
Well, glad to have helped :)
cold fossil
Collecting parselmouth
Using cached parselmouth-1.1.1.tar.gz (33 kB)
Preparing metadata (setup.py) ... done
Collecting googleads==3.8.0 (from parselmouth)
Using cached googleads-3.8.0.tar.gz (23 kB)
Preparing metadata (setup.py) ... error
error: subprocess-exited-with-error
× python setup.py egg_info did not run successfully.
│ exit code: 1
╰─> [1 lines of output]
error in googleads setup command: use_2to3 is invalid.
[end of output]
note: This error originates from a subprocess, and is likely not a problem with pip.
error: metadata-generation-failed
× Encountered error while generating package metadata.
╰─> See above for output.
note: This is an issue with the package mentioned above, not pip.
hint: See above for details.
any help
back (most recent call last):
File "C:\Users\ashwa\Desktop\RVC-GUI-main\rvcgui.py", line 23, in <module>
from vc_infer_pipeline import VC
File "C:\Users\ashwa\Desktop\RVC-GUI-main\vc_infer_pipeline.py", line 1, in <module>
import numpy as np, parselmouth, torch, pdb
ModuleNotFoundError: No module named 'parselmouth'
azure marshBOT
proper shale
^^ applio is a good bet
pine flame
Does yelling ruin the training in an AI model like a character saying "Whaaaaaaat!?"
proper shale
Does yelling ruin the training in an AI model like a character saying "Whaaaaaaa...
As long as it isn't frequent, no
sand iris
Don't think you'll get anything usable out of it
check pins in #🔍│help-w-okada , you can run this on google servers via google colab.
Not sure how you'd get the sound hooked up to voicechange though
molten pecan
THANK YOU 🙏
real igloo
guys please what does this mean: call, so the stacktrace below might be incorrect.
For debugging consider passing CUDA_LAUNCH_BLOCKING=1.
Compile with TORCH_USE_CUDA_DSA to enable device-side assertions.
i keep getting stuck on this whenever i try to use feature extraction 😭
and i can't train the model or the index
sand iris
guys please what does this mean: call, so the stacktrace below might be incorrec...
need the full error
molten pecan
does the reverb sound too extreme?
I don't hear any reverb
Also lower the breath sounds and search feature ratio a lil bit
https://aihubdocs.github.io/en/ check the inference settings page
keen snow
Also lower the breath sounds and search feature ratio a lil bit
aah ok tysm !!!
molten pecan
yw!
rare gobletBOT
Ayo? @strong girder level 1 !!!
molten pecan
yes
vast forge
where is this folder?
half cove
where is this folder?
thats a batch file and its located in your rvc folder
vast forge
thats a batch file and its located in your rvc folder
is it in the AMD version?
half cove
is it in the AMD version?
on amd one it would be names go-web-dml
vast forge
i downloaded it
sand iris
on amd one it would be names go-web-dml
they removed it
it's borked when you try to use it anyways
half cove
rips
vast forge
it worked
sand iris
amd inference is gone 🦀
vast forge
sand iris
Yeah try to infer anything
vast forge
what do i go with the dataset?
half cove
what do i go with the dataset?
your trying to train on amd?
vast forge
yes
half cove
vast forge
is it this?
rare gobletBOT
Ayo? @vast forge level 2 !!!
vast forge
half cove
you should use colab instead
vast forge
how do i do that
half cove
cuz ur on amd
half cove
it says in the guide...
vast forge
oh
half cove
after training model you get pth
after training index you get index
it should explain it all
half cove
for the dataset is it already index and pth
the dataset is your audios
your training a model right?
vast forge
im trying to make a model
half cove
how long
vast forge
about 8:46
half cove
how clean is the dataset
vast forge
clean enough i guess
half cove
should work then
vast forge
i followed exactly what the guide said
half cove
i followed exactly what the guide said
then just continue
vast forge
half cove
make a zipped folder
name it what u want (make sure no spaces or weird letters)
put the audio inside
then put in your disconnected folder in google drive
half cove
that works aswell
vast forge
no
half cove
what?
vast forge
where does it say in the guide to run it?
placid holly
can anyone help me how to calculate save epoch step?
i want my save epoch same as that Tensorboard show, like 770 step, i want my save to s770 too
like how many sample / batch i need to get exact number
vast forge
im going to run it
half cove
where does it say in the guide to run it?
scroll down and see...
vast forge
just tell me