#✨│ai-help
pastel oak
Then select the virtual cable as as one of the hardware inputs and activate A1 if your headphones are set to A1
im off to work again, ill try to check the chat if any other questions arise when i can
lament furnace
3 hours 
rare gobletBOT
Ayo? @lament furnace level 3 !!! 
fathom reef
Alright. Thanks for the help!
rare gobletBOT
Ayo? @fathom reef level 7 !!!
azure marshBOT
RVC Guides (How to Make AI Cover)
fathom reef
im off to work again, ill try to check the chat if any other questions arise whe...
Wait, hold on. What if other applications are using up a significant amount of CPU? Because when I switch to VRChat, my AI's voice lags back.
hazy thunder
lower batch size = higher quality?
fathom reef
Well yeah, but I think the RVC voice changer's using my CPU instead of GPU.
hazy thunder
i noticed my models which were trained on a lower quality dataset but on a batch size of 4 have higher quality than the one im training right now on batch size 8
fathom reef
oh wait
pastel oak
Wait, hold on. What if other applications are using up a significant amount of C...
the rvc-realtime doesnt use up much cpu fortunately. however, you will have to reduce ingame quality by a significant amount if it does reach the higher ends
fathom reef
the rvc-realtime doesnt use up much cpu fortunately. however, you will have to r...
I--
hazy thunder
lower batch size = higher quality?
could someone answer this bruh
pastel oak
Yes I was giving you a comparison to wokada for the cpu usage
send a screenshot of rvc
pastel oak
lower batch size = higher quality?
that is usually not the norm, but on shorter dataset of max 5 minutes it might be better to train on size 4 people say
fathom reef
hazy thunder
that is usually not the norm, but on shorter dataset of max 5 minutes it might b...
interesting, i mightve overtrained my model for it to sound bad then
pastel oak
alright nah everything looks decent, you can mess around with extra inference to lower it and see if the cpu usage goes down, but most likely wont. on the other hand, you can increase it if it doesnt cause higher cpu usage
pastel oak
interesting, i mightve overtrained my model for it to sound bad then
you using tensorboard to keep track?
fathom reef
No changes, sadly.
hazy thunder
you using tensorboard to keep track?
yea i am
fathom reef
Using this in VRChat is kinda funny, though. I can hear myself going like...
hazy thunder
turns out i overtrained by A LOT
hazy thunder
500 epochs for like a 30 minute dataset 😭
fathom reef
Still a problem, but funny nonetheless.
pastel oak
Hel-l-l-l-l-l-l-o
yea that means your cpu cant keep up with the sample length
youd need to go higher on it
fathom reef
Higher??
pastel oak
yea, higher means more delay
hazy thunder
gonna try 280 epochs cause on the graph the lowest point was like on epoch 270
fathom reef
Then why did it do the exact opposite of what you described??
pastel oak
gonna try 280 epochs cause on the graph the lowest point was like on epoch 270
Sounds good try that one
hazy thunder
yea im reloading the checkpoint
pastel oak
Then why did it do the exact opposite of what you described??
Lol it might take a few more seconds to stabilize after changing it
fathom reef
Uh, yeah. That Inference time is pretty bad...
Plus it didn't do anything to stop the glitching.
pastel oak
its to be expected on cpu, you really cant run this properly
fathom reef
Well, at least not much.
pastel oak
Well, at least not much.
well its as best as it gets for now. if i manage to figure out a better optimized way to run this on cpu ill let you know (yesterday it was shown how to do it but i havent gotten it to work on my amd based system so until i figure out what went wrong i cant help. after that i can pass it on to you)
dire vapor
"force_mbkm" What does this option do?
latent kettle
which one is good and for which purpose i can use them all. Pre-trained :: Orignal, Rin_E3, itaila, SnowieRU
carmine scaffold
why i cant post screenshots
anyway can someone help with this?
NameError Traceback (most recent call last)
<ipython-input-16-cc3a563dfbef> in <cell line: 27>()
25
26 pretrained_base = "pretrained/" if model_architecture == "v1" else "pretrained_v2/"
---> 27 unpt = f"_{pretrain_type}" if pretrain_type!="OV2" else ""
28
29 pretrainedD = f"{pretrained_base}f0D{target_sample_rate}{unpt}.pth"
NameError: name 'pretrain_type' is not defined
this comes when i click Training button
distant path
voice ai isnt opening
should i delete an reopen
*delete-redownload
also can i use voice ai in scp secret lab
rare gobletBOT
Ayo? @distant path level 1 !!!
distant path
im planning to troll
finite galleon
*delete-redownload
Delete and never use again
Use w okada voice changer
Fully free
carmine scaffold
what pretrain_type: should use
rare gobletBOT
Ayo? @carmine scaffold level 1 !!!
distant path
voice ai stuck on 90 percen tloading in app
polar crest
is there a google colabs rvc notebook that can let me train a model then resume training later?
dire ember
can anyone tell me how do i create a train voice model
turbid pagoda
Would anyone be able to tell me how to make a data set from plain vocals. And I've also got about 30 mins of audio to train. Can applio or llaria train that long?
I can't really remember how to make the data set with flac audio files
distant path
does w okada support custom voice models
for example like scp 079
or clanker droids
low shard
does w okada support custom voice models
Yes, btw ask in #🔍│help-w-okada or #1192011222023950368 for wokada related help/error/questions
distant path
so i can be b1 battle droid?
rustic aurora
hi
how can i make an ai model
is there any documentation
or is there any information about RVC
brittle wing
can anyone dm me or ping me the latest rvc link
rare gobletBOT
Ayo? @brittle wing level 2 !!!
pastel oak
how can i make an ai model
check pinned message, open the docs aihub guide, it has everything you need
rustic aurora
okay thx
hexed lynx
How do i convert other peoples voice models to ONXX? I am using an AMD GPU and the bot just said: "Don't forget that your models needs to be converted in ONNX!"
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 💖
desert niche
how do I change the pitch when inferencing using gpt sovits? I trained a female voice but when I use it, a male voice comes out
is this because I used a male voice audio?
Idk if I should ask here or in #🔊│ai-development
@half cove u know anything about gpt sovits?
granite ravine
how do i import onnx filees
half cove
<@936673139419664414> u know anything about gpt sovits?
i dont think it has pitch
maybe depends on reference audio
or you can use a daw
azure marshBOT
RVC Guides (How to Make AI Cover)
half cove
change the pitch in a daw?
yep
i mean in gptsovits its not singing so pitch isnt needed much
and emotion is based on reference audio
rare gobletBOT
Ayo? @kind sentinel level 1 !!!
kind sentinel
Then I PAID for credits
And still can't make my covers by an error 😭😭
The thing is that I already paid
For 100 credits
wicked arrow
why does my res go up to like 20k
kind sentinel
Nvm, I already can
errant wyvern
I need help, I installed cuda toolkit 12.4 and change voice stopped working.
the software gives this error. Web Server Launch Exception, D:\a_work\1\s\onnxruntime\python\onnxruntime_pybind_state.cc:574 onnxruntime::python::CreateExecutionProviderInstance CUDA_PATH is set but CUDA wasn't able to be loaded. Please install the correct version of CUDA and cuDNN as mentioned in the GPU requirements page (https://onnxruntime.ai/docs/reference/execution-providers/CUDA-ExecutionProvider.html#requirements), make sure they're in the PATH , and that your GPU is supported.
What do I need to install?
upper furnace
frail fox
yo guys i got question what to do if they told you the voice isn't suctesfully instalized, please try again. should ichange it in option or what?
if i actually instalized what i seen in tutorial
shouldi unzip folder or something?
errant wyvern
I managed to get it to work, I believe some file was corrupted or something. because I just downloaded it again and extracted it and it worked.
rare gobletBOT
Ayo? @errant wyvern level 1 !!!
frail fox
I managed to get it to work, I believe some file was corrupted or something. bec...
but i don't have enought coins for train it again
errant wyvern
but i don't have enought coins for train it again
???
@frail fox what are you talking? I didn't understand.
frail fox
???
you know i mean this credits
my character isn't working but anyone who i pick isn't working
foggy folio
-uvr
azure marshBOT
-uvr
- Perfecting Audio Isolation on Low-End Rigs: A Practical Guide, by Litsa The Dancer and Faze Masta Google Docs
- Gathering and Isolating Audio, by SCRFilms Google Docs
- Creating Datasets for RVC using iZotope RX, by LUSBERT
Rentry - Vocal Mixing Tutorial, by Roomie YouTube
stiff goblet
Guys i have a small dataset around 3 minutes. It's not in English language. Which pretrain model should I use ? OV2 or original ?
azure marshBOT
austere mortar
where do i use a .ckpt?
short lintel
what is the index file?
low shard
you know i mean this credits
Credits? Are you using voice.ai or another platform?
frail fox
Credits? Are you using voice.ai or another platform?
it were working one'c upon a time
but now
not any of the voices are working
slender laurel
do u need the coin things for voiceai
low shard
voice.ai
We aren't affiliated, it's paid, we use rvc, which is all free and you dont use credits, you cant upload voice.ai models here, all voice models here are either rvc or gpt so vits (which is another thing also free of tts but for ai covers you van use rvc)
low shard
do u need the coin things for voiceai
Yea its paid lol, its better to use rvc which is open source and no need to pay coins for it lol
frail fox
Yea its paid lol, its better to use rvc which is open source and no need to pay ...
what is this "rvc" thing?
slender laurel
Yea its paid lol, its better to use rvc which is open source and no need to pay ...
are rvc and w okada the same
low shard
what is this "rvc" thing?
Its the free tool for voice conversion (longer said as retrevial-based voice conversion), aka ai covers tool, if you want to know how to use it you could check out our docs https://docs.aihub.wtf
low shard
are rvc and w okada the same
Nope, the difference is wokada is a realtime voice changer (it uses rvc models too, so u can use the ones from the server dw) rvc instead converts from an audio file to another
slender laurel
Nope, the difference is wokada is a realtime voice changer (it uses rvc models t...
o i thought rvc was short for realtime voice changer 💀
low shard
o i thought rvc was short for realtime voice changer 💀
Its short for Retrevial-based Voice Conversion lol
short lintel
hey im using the applio rvc for the first time, and i cant seem to get it to work by just putting the voice model link from the discord. any help what am i doing wrong, thanks!
low shard
hey im using the applio rvc for the first time, and i cant seem to get it to wor...
Send the model download link
low shard
https://huggingface.co/AIVER-SE/LanaDelRey/resolve/main/LanaDelRey.zip
Mmm okay the model is fine, are you using applio locally or the colab?
short lintel
colab
so i should just be able to copy the link into the voice model tab click upload voice and it should be there?
low shard
so i should just be able to copy the link into the voice model tab click upload ...
Yea after you done that you should click refresh
And it should appear there
short lintel
really dont know what i could be doing wrong, what is the current link to the colab page i should be using?
rare gobletBOT
Ayo? @short lintel level 2 !!!
low shard
really dont know what i could be doing wrong, what is the current link to the co...
if you wanna use applio then this https://colab.research.google.com/github/iahispano/applio/blob/master/assets/Applio.ipynb
do you get any errors downloading the model ?
or it says the model download fine ?
short lintel
in the interface? no it just loaded and dissapeared
low shard
in the interface? no it just loaded and dissapeared
Dissapeared? Retry to download it againand click refresh and see if its there, if its not send me screenshots of the colab output cell
rugged stirrup
I'm getting this error when training index RuntimeError: Error in __cdecl faiss::FileIOWriter::FileIOWriter(const char *) at D:\a\faiss-wheels\faiss-wheels\faiss\faiss\impl\io.cpp:98: Error: 'f' failed: could not open logs/Ninomae-Ina’nis-HoloLive ENG/trained_IVF293_Flat_nprobe_1_Ninomae-Ina’nis-HoloLive ENG_v2.index for writing: No such file or directory
short lintel
omg.. as i said im new to this. wasn't downloading the voice just pasted the link int the upload voice tab.. smh. but thank you for your time brother, couldn't have figured it out on my own
shell crescent
hey, trying to make an ai cover using colab and in step 2 i have an error. new to the whole thing. any help will be appreciated.
severe sand
So if I understood correctly, I would want to pick a model at a step where the g loss was the lowest right?
What I don't get though is that the datapoints in tensorboard are at entirely different steps than the models I saved.
As example 8400 steps is supposedly the lowest point, but I only have a models at steps 8160 and 9180.
pastel oak
So if I understood correctly, I would want to pick a model at a step where the g...
You just take the one closest to before the lowest
severe sand
so the 8160 in my case, okay then.
How are the datapoints for tensorboard written? is that in control of RVC or tensorboard?
I mean is it tensorboard only showing every 200 steps or is that RVC writing down every 200 steps.
sage oar
Question: Can i use RVC models on android?
And if so,how?
severe sand
Question: Can i use RVC models on android?
And if so,how?
I don't think the usual android phone has the required hardware for it, I could be wrong though
sage oar
Well crap
Jdk either💀
I was writing and translating some lyrics from a song
And i realized that i could use ai voice changer to like
Change the tone to the character singing the song
So yeah
Just to test out the lyrics
rare gobletBOT
Ayo? @sage oar level 1 !!!
sage oar
Just to test out the lyrics
Thats why im asking
severe sand
Well if you have it running on a computer in the same network you could access it through the phones browser
sage oar
I wish
I dont exactly have the fastest computer
Were talking like windows 8 type-
😭😭😭💀
Its REALLY SLOW
pastel oak
Thats why im asking
ive seen people train voice models through colab on their phone so im sure you could use rvc with that
but for the realtime ai voice changer its not gonna work
pastel oak
so the 8160 in my case, okay then.
How are the datapoints for tensorboard writt...
Not sure i think thats just normal
lean bane
everytime i try to open applio's tensorboard i get "No path found after logs" error on repeat. How do i fix
gray plaza
little ridge
anybody can help me figure out why I keep getting this message when i input an AI Model?
rare gobletBOT
Ayo? @little ridge level 1 !!!
little ridge
JSONDecodeError Traceback (most recent call last)
<ipython-input-7-c3a94fc37506> in <cell line: 31>()
31 if os.path.exists(config_path):
32 # File exists, proceed with creation of creds and client
---> 33 creds = Credentials.from_service_account_file(config_path, scopes=scope)
34 client = gspread.authorize(creds)
35 else:
5 frames
/usr/lib/python3.10/json/decoder.py in raw_decode(self, s, idx)
353 obj, end = self.scan_once(s, idx)
354 except StopIteration as err:
--> 355 raise JSONDecodeError("Expecting value", s, err.value) from None
356 return obj, end
JSONDecodeError: Expecting value: line 1 column 1 (char 0)
thin stump
---------------------------------------------------------------------------
JSON...
You're using an outdated colab so check here https://docs.aihub.wtf/
little ridge
Thank you!
proper shale
not sure, at least in replay
brittle wing
If I resume the current session if I only run load preprocessed dataset files from Google drive,import model from drive to notebook & training cells the target sample rate I'm training on is memorized right???
delicate wedge
Is weights down?
brittle wing
Is weights down?
Weightsgg?
delicate wedge
Yeah
delicate wedge
Oh
Apparently I got banned.
I made two rvc models today.
I'm not sure if they will were against terms but damn.
Oh nvm it's working again lol
pale thunder
i asked this earlier, idk if i should move to w-okada or not. any RVC model i use doesnt work
rare gobletBOT
Ayo? @pale thunder level 1 !!!
azure marshBOT
RVC Guides (How to Make AI Cover)
unique cave
hay guys does colab work on mac?
proper shale
hay guys does colab work on mac?
yeah it's a browser application
works on quite literally anything that has an internet connection
gleaming kettle
it keeps giving me this err JSONDecodeError Traceback (most recent call last)
<ipython-input-12-2c22fc1fda04> in <cell line: 31>()
31 if os.path.exists(config_path):
32 # File exists, proceed with creation of creds and client
---> 33 creds = Credentials.from_service_account_file(config_path, scopes=scope)
34 client = gspread.authorize(creds)
35 else:
5 frames
/usr/lib/python3.10/json/decoder.py in raw_decode(self, s, idx)
353 obj, end = self.scan_once(s, idx)
354 except StopIteration as err:
--> 355 raise JSONDecodeError("Expecting value", s, err.value) from None
356 return obj, end
JSONDecodeError: Expecting value: line 1 column 1 (char 0)
rugged stirrup
What should I do with a spectrogram that looks like this?
bronze maple
What should I do with a spectrogram that looks like this?
your sample indicates that it is at 44000Hz, this is what you must check on the RVc application to train a model
young oasis
what's the best extraction method for converting from voice to voice?
neat briar
hey uhm
question
how do i use the voice for things like whatsap web or discord and etc
mellow token
neat briar
but what exactly , theres a lot
rare gobletBOT
Ayo? @neat briar level 1 !!!
umbral falcon
im getting an error when i try to use ilaria rvc it tries to process then immediately says error help please 😄
wanton bear
yo how tf u fix the 0 vol issue
stray adder
Can someone send me a webpage to change the voice? Google Collab is getting error
tawdry wind
Anyone have any suggestions on how I can clean up the vocals on a song for an AI cover? I want to try doing one of Tommy the Cat by Primus, but the fast vocals and the radio effect I think is muddying the results 🤔
stray adder
Still getting this error message
tawdry wind
Last i heard everyone was getting that error
haughty moss
train compleate but notebook to drive , but does not show .pth file only showing index
finite galleon
Still getting this error message
Thats an outdated colab and may i ask where did you get that link from?
haughty moss
can not find pth file
brittle wing
yeah so i have voice ai and im trying to get a voice but like what do i do with the thing. like i search the desired ai voice and then its just a gif and i dont know what to do with it
livid frigate
Is there a way to make like the ai voice got some emotion too without some mp3?
spare gull
Can I stop training suddenly?
rare gobletBOT
Ayo? @low lynx level 2 !!!
lament eagle
What is this problem in RVCDisconnected
azure marshBOT
RVC Guides (How to Make AI Cover)
summer glacier
anyone got an app that i can use to test the voices i made?
spare gull
Rvc
summer glacier
thats the creation thing
im trying to fix a replay error
import scipy
oops
Error creating song: No module named 'scipy.special._cdflib'
stiff goblet
guys why i'm getting this error in tensorboard
hoary crown
I keep getting the Error "32 # File exists, proceed with creation of creds and client"
It also shows up as "JSONDecodeError: Expecting value: line 1 column 1 (char 0)"
What do I need to do to fix it?
hoary crown
https://discord.com/channels/1159260121998827560/1159290139609137264/12252261922...
Thank You
short lintel
how do you guys get the cleanest vocals for conversion? UVR?
thin stump
how do you guys get the cleanest vocals for conversion? UVR?
https://mvsep.com/en BS Roformer finetuned Viperx
It will better than MDX23 on UVR
and the beta version of Roformer on uvr
novel elbow
hello?How to get start to creat a model voice?
proper shale
hello?How to get start to creat a model voice?
read https://docs.aihub.wtf/ ;)
rugged stirrup
Why is the RVC quality control bot thing removing my models for pretty much no reason for "low quality" even though I've seen other low quality models that don't get removed
rare gobletBOT
Ayo? @rugged stirrup level 22 !!!
azure marshBOT
RVC Guides (How to Make AI Cover)
spare gull
How to stop training
I already train 145
I wanna stop
But I set train 200
How to stop and save the training
short lintel
can a model trained on for example english, can convert and/or tts, in different languages?
crude dome
Ugh, hi there
I might have a problem with ai...
sounddevice.PortAudioError: Error opening Stream: Illegal combination of I/O devices [PaErrorCode -9993]
Did you know what's wrong?
sand iris
sounddevice.PortAudioError: Error opening Stream: Illegal combination of I/O dev...
Make sure both audio devices are on the same protocol
It should be between parentheses
crude dome
If only I knew what are you talking about
rare gobletBOT
Ayo? @crude dome level 1 !!!
sand iris
If only I knew what are you talking about
Share a screenie
crude dome
U mean in voice channel?
sand iris
Send a screenshot of the program here
crude dome
sand iris
the other part of the program
crude dome
ocean socket
-colab
solid mango
hello
i need a help
i get this error
ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts.
chex 0.1.86 requires numpy>=1.24.1, but you have numpy 1.23.5 which is incompatible.
pandas-stubs 2.0.3.230814 requires numpy>=1.25.0; python_version >= "3.9", but you have numpy 1.23.5 which is incompatible.
torchtext 0.17.1 requires torch==2.2.1, but you have torch 2.0.1 which is incompatible.
torchvision 0.17.1+cu121 requires torch==2.2.1, but you have torch 2.0.1 which is incompatible.
idk what to do with it
its on colab
spare gull
after i open the realtime gui
rare gobletBOT
Ayo? @spare gull level 3 !!!
waxen rover
Heya, sorry if this is an often-asked question, but why does training restart even though the model was loaded successfully? am i missing files? (Both d and g files and the config.json were under drive/rvcDisconnected/modelname)
proper shale
Heya, sorry if this is an often-asked question, but why does training restart ev...
you also need the rvcLogs.zip
and you also need to make sure the settings (batch_size and the training variables) are the same
waxen rover
you also need the rvcLogs.zip
ahh that sucks
where do i download it next time i want to quit and resume training?
waxen rover
oh wait yeah i did also import those, that's the preprocessed dataset files one right
rare gobletBOT
Ayo? @waxen rover level 1 !!!
brittle wing
my res is 12k is that supposed to happen? 😭
spare gull
after u run "Save preprocessed files to drive"
sir,my voice changer gui crashed after i open my mic
is there any problem about my file?
brittle wing
my res is 12k is that supposed to happen? 😭
well its 20k rn
brittle wing
after u run "Save preprocessed files to drive"
do seem to know?
btw im using the local one
proper shale
oh wait yeah i did also import those, that's the preprocessed dataset files one ...
yeah
proper shale
maybe it didn't properly save the d/g files? had that happen once
when u stop in the middle of saving d/g files on a saving epoch
waxen rover
i'll try loading an earlier one
brittle wing
@proper shale pls help
rare gobletBOT
Ayo? @brittle wing level 1 !!!
brittle wing
20k res
proper shale
btw im using the local one
check https://rentry.co/VoiceChangerGuide
if you're using amd, ya need to convert downloaded models to onnx, guide explains it
proper shale
sir,my voice changer gui crashed after i open my mic
not sure how to deal with that tbh
brittle wing
check https://rentry.co/VoiceChangerGuide
if you're using amd, ya need to conve...
i used nvidia and i downloaded nvidia
but its so laggy
maybe i have a low end pc
is cloning smoother for low end? @proper shale
proper shale
maybe i have a low end pc
what's your GPU
proper shale
it just said python have no response
again, not sure how to deal with that
brittle wing
what's your GPU
gtx 950m 4 gb 😭
proper shale
sorry
proper shale
gtx 950m 4 gb 😭
yeah that's probably why
proper shale
you should use colab for rvc stuff:
brittle wing
you should use colab for rvc stuff:
https://colab.research.google.com/github/hi...
oh
i used the illaria
is that ok too
spare gull
btw,i got my index and pth file,but i can't upload on my rvc
brittle wing
@proper shale
proper shale
i used the illaria
if you're using colab yes
brittle wing
ok thx
proper shale
btw,i got my index and pth file,but i can't upload on my rvc
you should use W-Okada if you can
spare gull
you should use W-Okada if you can
what is W-Okada
latent kettle
btw,i got my index and pth file,but i can't upload on my rvc
are you trying to uplaod g_2333333
proper shale
just saying
proper shale
what is W-Okada
the realtime voice changer that uses rvc
latent kettle
what is W-Okada
realtime voice changer
spare gull
may i get the file
brittle wing
@proper shale well uh what character should i use
spare gull
is this W okada
brittle wing
@proper shale what is audio file how to get
spare gull
is this W okada
@proper shale
proper shale
you a recommendation?
dont got anything
proper shale
is this W okada
yes
proper shale
<@622856015444049937> what is audio file how to get
an .mp3?
look, read https://docs.aihub.wtf/ if you're using anything other than W-Okada
brittle wing
oh.
spare gull
i cant upload the file
brittle wing
@proper shale i couldnt use the realtime?
brittle wing
man
rare gobletBOT
Ayo? @brittle wing level 2 !!!
proper shale
i cant upload the file
make sure you're uploading the file from the weights folder, if you've trained it
proper shale
that my pc can use :>
nope, sadly
proper shale
yeah
proper shale
this is not relevant towards the voice changers
you need a good gpu
at least a 1650
proper shale
maybe try to upload the pth only?
spare gull
maybe try to upload the pth only?
can u try to upload
stable barn
Hello Everyone. Does anyone know here how to use the Realm AI Chrome Extension?
spare gull
might be an issue on your end? maybe try a fresh reinstall of W-Okada
It shows config missing
What does it mean
prime talon
How do people separate the harmonies in the acapella for Bohemian Rhapsody covers?
rare gobletBOT
Ayo? @prime talon level 1 !!!
waxen rover
Heya, sorry if this is an often-asked question, but why does training restart ev...
i really don't understand what i could be doing wrong :') been trying to fix it for an hour now
proper shale
How do people separate the harmonies in the acapella for Bohemian Rhapsody cover...
might be using the original multitracks
proper shale
how to fix
honestly idk. might be a model issue
spare gull
Pls anyone help
cunning void
how can ı see the voice models
fresh mantle
This doesn’t work for gradio
silk pasture
can anyone help with settings or just the place where the problem can be, please? I downloaded rvc successfully and put my microfone and speakers but still after start and saying something I don't get any output voice.. Have no idea where the problem can be😅
azure marshBOT
RVC Guides (How to Make AI Cover)
echo pecan
-colab
fresh mantle
I need help im trying to get minji voice model but it doesn’t work for gradio live
stable oak
Idk if this is dumb but I have a question, is there a way to do text to speech with one of the voice models I downloaded?
fiery raptor
why am i getting this error
proper shale
why am i getting this error
model link is screwed up, try another model ig
proper shale
Idk if this is dumb but I have a question, is there a way to do text to speech w...
not natively, but you can use tts tools like ElevenLabs and then convert the result with RVC
stable oak
not natively, but you can use tts tools like ElevenLabs and then convert the res...
I LOVE YOU THANKS
proper shale
yw :)
proper shale
what are ya using
cursive condor
only works the "beatrice"
proper shale
oh, W-Okada
cursive condor
ye
proper shale
check cmd, switch to an rvc model and see if any Traceback pops up
cursive condor
well, i have a rtx 4070 and a amd 7600x
rare gobletBOT
Ayo? @cursive condor level 1 !!!
graceful basalt
hello can I ask something?
cursive condor
I don't understand why it doesn't work
cursive condor
check cmd, switch to an rvc model and see if any Traceback pops up
nope, only works beatrice
both the rvc that were already in the application and the ones that I added do not work
proper shale
nope, only works beatrice
I'm aware, but I'm asking you to check command line
Not to see if it works or not (because it seems to not work)
proper shale
So apparently something went wrong with the initial install once you open W-Okada. Try deleting the pretrain and model_dir folders (inside MMVCServerSIO) and restarting the program. It should re-do that initial installation and should work after that. Just make sure to keep the cmd open until the installation ends
cursive condor
K
proper shale
yw :)
blazing jungle
hi
i had a problem executing TB
Launching TensorBoard...
Traceback (most recent call last):
File "<frozen runpy>", line 198, in _run_module_as_main
File "<frozen runpy>", line 88, in run_code
File "C:\Users\goodl\AppData\Local\Programs\Python\Python311\Scripts\tensorboard.exe_main.py", line 4, in <module>
File "C:\Users\goodl\AppData\Local\Programs\Python\Python311\Lib\site-packages\tensorboard\main.py", line 40, in <module>
from tensorboard import default
File "C:\Users\goodl\AppData\Local\Programs\Python\Python311\Lib\site-packages\tensorboard\default.py", line 38, in <module>
from tensorboard.plugins.audio import audio_plugin
File "C:\Users\goodl\AppData\Local\Programs\Python\Python311\Lib\site-packages\tensorboard\plugins\audio\audio_plugin.py", line 25, in <module>
from tensorboard import plugin_util
File "C:\Users\goodl\AppData\Local\Programs\Python\Python311\Lib\site-packages\tensorboard\plugin_util.py", line 21, in <module>
from tensorboard.vendor import bleach
File "C:\Users\goodl\AppData\Local\Programs\Python\Python311\Lib\site-packages\tensorboard_vendor\bleach_init.py", line 5, in <module>
from tensorboard._vendor.bleach.linkifier import (
File "C:\Users\goodl\AppData\Local\Programs\Python\Python311\Lib\site-packages\tensorboard_vendor\bleach\linkifier.py", line 4, in <module>
from tensorboard.vendor import html5lib
File "C:\Users\goodl\AppData\Local\Programs\Python\Python311\Lib\site-packages\tensorboard_vendor\html5lib_init.py", line 25, in <module>
from .html5parser import HTMLParser, parse, parseFragment
File "C:\Users\goodl\AppData\Local\Programs\Python\Python311\Lib\site-packages\tensorboard_vendor\html5lib\html5parser.py", line 8, in <module>
from . import _tokenizer
File "C:\Users\goodl\AppData\Local\Programs\Python\Python311\Lib\site-packages\tensorboard_vendor\html5lib_tokenizer.py", line 16, in <module>
from .trie import Trie
File "C:\Users\goodl\AppData\Local\Programs\Python\Python311\Lib\site-packages\tensorboard_vendor\html5lib_trie_init.py", line 3, in <module>
from .py import Trie as PyTrie
File "C:\Users\goodl\AppData\Local\Programs\Python\Python311\Lib\site-packages\tensorboard_vendor\html5lib_trie\py.py", line 6, in <module>
from .base import Trie as ABCTrie
File "C:\Users\goodl\AppData\Local\Programs\Python\Python311\Lib\site-packages\tensorboard_vendor\html5lib_trie_base.py", line 3, in <module>
from collections import Mapping
ImportError: cannot import name 'Mapping' from 'collections' (C:\Users\goodl\AppData\Local\Programs\Python\Python311\Lib\collections_init.py)
Keeping the command prompt open...
Для продолжения нажмите любую клавишу . . .
what do i do?
lean agate
🩷 someone that uses rvc1006nvidia could help me? The program randomly lags, for no reason and when everything should be working well. I just open the settings of the pc, literally the settings of windows, and it lags for few secons. And is even worse in games. Idk why that happens if the gpu is in a normal temperature. it looks like is something else so pls if you can help me I would be so thanksful 🙏🙏🌸
scarlet rover
can someone link me the latest nvidia rvc? thanks in advance
proper shale
🩷 someone that uses **rvc1006nvidia** could help me? The program randomly lags...
alr answered ya ;)
proper shale
can someone link me the latest nvidia rvc? thanks in advance
mainline or do you want the realtime voice changer
scarlet rover
mainline or do you want the realtime voice changer
realtime please
brave garnetBOT
⠀
Download for Nvidia GPUs
Version 18a cuda
Download for AMD GPUs 
Version 18a directml
Download for Intel GPUs 
Version 18a directml
Download for Mac 
Version 17b Mac
⠀
Made with love by Ilaria 💖
proper shale
^ ;)
rugged stirrup
What sample rate should this be?
proper shale
What sample rate should this be?
32k
wispy burrow
What sample rate should this be?
What's this spectogram for?
There doesn't seem to be a single frequency 20kHz.
The highest frequencies are probably somewhere 18KHz.
So I would say 32Khz for the Sample Rate too.
rugged stirrup
I didnt make the dataset btw
stiff goblet
Reason of error? I'm using Mangio-Fork @proper shale
rare gobletBOT
Ayo? @stiff goblet level 6 !!!
echo wing
when i attempt to use the ai voice it sounds exactly like my normal voice and doesnt work
barren fern
Hey guys! Is there any googleColab RVC link that works? I haven't used it since the very old version of it and now I'm so confused with these new updates
long stirrup
Hi back guys,
I want to try my first AI model, but RVC is giving me error when i try to convert
i have my index and .pth, idk where i failed 
sour creek
What does that mean?
long stirrup
What does that mean?
maybe just try to zip you folder with winrar ?
long stirrup
click on your folder, you should have a line 'Compress Files to zip'
sour creek
Doesn’t show here
long stirrup
compress
sour creek
Oh okay
long stirrup
i'ts suppose to create you a .zip
wispy burrow
Hi back guys,
I want to try my first AI model, but RVC is giving me error when ...
Maybe the path for your audio file is wrong?
long stirrup
the path look good 🤔
wispy burrow
Oh, French.
yeah I think the path is wrong.
long stirrup
i tried same with Mangio i have the same error in console
the audio i want to convert is supposed to be somewhere specifically ?
wispy burrow
Right-Click the audio file and select "Copy as path."
Then paste that in RVC.
remove the ""
wispy burrow
the audio i want to convert is supposed to be somewhere specifically ?
It doesn't.
But you can place your audio files in the audios folder.
Your path has to end with TestThibaultv1.wav
or mp3 or whatever your audio format is
long stirrup
it works only when i put my files on desktop, the path is not found when it's on a folder, (surely I am not copying the path correctly)
wispy burrow
What’s that?
What is in your zip?
long stirrup
btw it workI 'm very grateful @wispy burrow
sour creek
What is in your zip?
Rev.mp3.zip
wispy burrow
Rev.mp3.zip
You need to have your model's. pth & .INDEX files.
sour creek
How do I get does?
wispy burrow
How do I get does?
you can only get by training your Model.
sour creek
How do I train my model?
long stirrup
https://docs.aihub.wtf/ @sour creek i just learnt how to train a model here, the tutorial is insane
sour creek
Okay thanks
wispy burrow
https://docs.aihub.wtf/ <@712973358114799686> i just learnt how to train a model...
You beat me to it!
I was searching for that one.
prime talon
might be using the original multitracks
Thanks! This was it, someone ripped stems that separated every layer of vocals and made it available publicly
rugged stirrup
When trying to open tensorboard on Ilaria RVC this happens could anyone help fix this?
azure marshBOT
RVC Guides (How to Make AI Cover)
rugged stirrup
I downloaded the Ov2 pretrain and its giving me this error when I train the model 2024-04-04 19:16:24 | INFO | __main__ | Use gpus: 0 2024-04-04 19:16:24 | INFO | __main__ | "runtime\python.exe" infer/modules/train/train.py -e "Boyfriend-FNF-Tails-Gets-Trolled" -sr OV2-32k -f0 1 -bs 6 -g 0 -te 140 -se 20 -pg assets/pretrained_v2\f0Ov2Super32kG.pth -pd assets/pretrained_v2\f0Ov2Super32kD.pth -l 1 -c 0 -sw 1 -v v2 Traceback (most recent call last): File "C:\Users\X3G9H\Downloads\Ilaria RVC Beta 1\infer\modules\train\train.py", line 14, in <module> hps = utils.get_hparams() File "C:\Users\X3G9H\Downloads\Ilaria RVC Beta 1\infer\lib\train\utils.py", line 368, in get_hparams config = json.load(f) File "json\__init__.py", line 293, in load File "json\__init__.py", line 346, in loads File "json\decoder.py", line 337, in decode File "json\decoder.py", line 355, in raw_decode json.decoder.JSONDecodeError: Expecting value: line 1 column 1 (char 0)
It works in colab but not locally
silk gyro
Hey guys, really dumb question. I have a 3070Ti already, and just got a GT 1030 exclusively for w-okada. Do I need to install any additional drivers? I don't think so right? Because my 3070Ti driver package should include everything already? Thanks in advance!
mellow juniper
im trying to use these files in voicemod and its saying it cannot download this sort of file any help?
nvm
formal elk
.
brittle wing
Hello, does anyone know how to make the AI voice heard in any game? It's just that I only listen to it
rare gobletBOT
Ayo? @brittle wing level 1 !!!
fleet marsh
hello can someone help me with my model? i tested my model and a old model i did 1 year ago
the old one sounds better
this is the old one
and this the new one
was the same dataset
thin stump
the old one sounds better
Might be one of those examples where the overtraining sounded better on the old one? We don't know what happened on the tensorboard except you
fleet marsh
Might be one of those examples where the overtraining sounded better on the old ...
the old one wasnt overtrained
maybe the new was overtrained
but it wasnt
thin stump
@drifting pike
drifting pike
hi
thin stump
hi
hey, and yeah that's pretty much it. Do you have the dataset saved where you didn't truncate yet (removing silences)
drifting pike
well i didnt really need to since most of the times when jack talked it was silent with no bg music
thin stump
It can train on white noise even if you're 100% sure there's nothing there
You can export it as an audio label
here i'll show you
drifting pike
ok
drifting pike
i see
thin stump
Then you just zip up that folder with the multiple files and save onto the rvc disconnected folder, simple 
drifting pike
oh i use applio locally
thin stump
oh same idea
drifting pike
oh ok
gritty radish
yall what do i download to make the ai covers not the models 😭 🙏
i had it before and i deleted it
thin stump
i had it before and i deleted it
Was it tryreplay.io?
gritty radish
nu uh
it was like this thing that used python
and it was an app right
it would pop up a blue rectangle kinda thing that asked for the model and like crepe length or smth
rugged stirrup
I downloaded the Ov2 pretrain and its giving me this error when I train the mode...
Could anyone help me with this?
drifting pike
But you did noise gate it right? It takes out some of the white noises
i kind of dont know how it works
rare gobletBOT
Ayo? @drifting pike level 2 !!!
thin stump
Could anyone help me with this?
I didn't mean to bury you 
. You should put in help-forum so it's more permanent
rugged stirrup
I didn't mean to bury you <:f_cry:1159653986681499768> . You should put in help-...
Oh your fine
gritty radish
Check here https://docs.aihub.wtf/essentials/how-to-make-ai-cover/
i have been for 10 minutes and its not the right one
thin stump
i kind of dont know how it works
Copy my noise gate settings here https://rentry.org/RVC-dataset-RX-updated or use the server guide https://docs.aihub.wtf/rvc/resources/datasets/#step-1-noise-gate
drifting pike
ok
winter ferry
shell gulch
glad falcon
What’s the best free app I can use of my pc for the voices?
somber gale
Hello, I'm trying to install Mangio right now andf ran into an issue while running "go-web.bat":
F:\RVC\Mangio-RVC-v23.7.0>runtime\python.exe infer-web.py --pycmd runtime\python.exe --port 7897
Traceback (most recent call last):
File "F:\RVC\Mangio-RVC-v23.7.0\infer-web.py", line 79, in <module>
conn = sqlite3.connect('TEMP/db:cachedb?mode=memory&cache=shared', check_same_thread=False)
sqlite3.OperationalError: unable to open database file
F:\RVC\Mangio-RVC-v23.7.0>pause
Press any key to continue . . .
hard cobalt
can i get some help?
twilit terrace
torch.cuda.OutOfMemoryError: CUDA out of memory. Tried to allocate 548.00 MiB (GPU 0; 6.00 GiB total capacity; 4.05 GiB already allocated; 0 bytes free; 5.29 GiB reserved in total by PyTorch) If reserved memory is >> allocated memory try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF
Traceback (most recent call last):
File "C:\Users\finnd\Desktop\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 "C:\Users\finnd\Desktop\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 "C:\Users\finnd\Desktop\Mangio-RVC-v23.7.0\runtime\lib\site-packages\gradio\blocks.py", line 1283, in postprocess_data
prediction_value = block.postprocess(prediction_value)
File "C:\Users\finnd\Desktop\Mangio-RVC-v23.7.0\runtime\lib\site-packages\gradio\components.py", line 2586, in postprocess
file_path = self.audio_to_temp_file(
File "C:\Users\finnd\Desktop\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'```
I'm aware it says I have no memory I'm just not sure how to change that let alone if thats the only problem. Any help is appricated :3Can sb help with this?
rare gobletBOT
Ayo? @somber gale level 1 !!!
twilit terrace
Can sb help with this?
damn by the looks of how no ones responded to you... I got little faith in anyone responding to me lol
somber gale
damn by the looks of how no ones responded to you... I got little faith in anyon...
😭
u got any idea how to fix it?
somber gale
```x_flat = F.pad(x_flat, commons.convert_pad_shape([[0, 0], [0, 0], [length, 0]...
try to free up somee space honestly thats the only thing i can think off
twilit terrace
ive tried clearing half of my hard drive but i suppose there stuff i can always reinstall
rare gobletBOT
Ayo? @twilit terrace level 1 !!!
somber gale
conn = sqlite3.connect('TEMP/db:cachedb?mode=memory&cache=shared', check_same_thread=False)
the db folder is inexistent in my temp folder
@twilit terrace can you Windows + R run "temp" and check if a folder called "db" exists?
twilit terrace
theres nothing
somber gale
i found t he issue
they said place it in C...
i placed it in F
im a retard
place it in C
place it in C
place it in C
twilit terrace
im a retard
dw i just realized you placed it in f too... istg bug fixing is like a wheres waldo game on steriods
either way i found my error too
my sample was too big
needed to cut it in half
twilit terrace
yeah ive always put it in C
i had another AI model thing that i cant seem to find anymore that I initially tried to install in F but it just fucked it up
somber gale
hmm
twilit terrace
didn't refuse to boot tho so ig it may be diff
somber gale
my sample was too big
makes sense
twilit terrace
im trynna copy a german youtubers voice
im making songs out of my principals voice
as a joke for a school thing
somber gale
im making songs out of my principals voice
LOL
rare gobletBOT
Ayo? @somber gale level 2 !!!
somber gale
thats crazy
somber gale
thank you, you too 😄
somber gale
thx for the help lol
no worries I just let chatgpt help me but it couldnt help on my issue 😦
twilit terrace
no worries I just let chatgpt help me but it couldnt help on my issue 😦
well get to a point some day when ChatGPT can help with that kind of stuff
rn its coding side really leaves something to be desired
somber gale
rn its coding side really leaves something to be desired
I agree but github copilot is great
brittle wing
Sorry guys to bother you, but how do I use the voice model I downloaded?
twilit terrace
I agree but github copilot is great
never used copilot I prolly should tho ive seen it in action before looks rad
rare gobletBOT
Ayo? @twilit terrace level 2 !!!
somber gale
Sorry guys to bother you, but how do I use the voice model I downloaded?
You trying to AI cover or voice change?
somber gale
never used copilot I prolly should tho ive seen it in action before looks rad
its goooog
twilit terrace
Sorry guys to bother you, but how do I use the voice model I downloaded?
sure we both came here lookin for help but ig i can see what i can do
brittle wing
voice change
twilit terrace
what software you using?
brittle wing
what software you using?
I thought it was on your cell phone
brittle wing
watch this
Ok
somber gale
you insert your model and index intop okada
old oracle
-colab
brittle wing
you insert your model and index intop okada
Ok thx
twilit terrace
works now
ggs!
now just dont make your input audio files too big or you end up like i did lol
twilit terrace
I just wanna train that thing in peace now with no problems
i know the feeling lol
dim zodiac
Any working training colab?
brittle wing
Does anyone know how hardware affects model training quality? I understand the different training settings, but will quality be affected if training is done on a bad GPU vs good GPU, besides time needed to complete training ofc. I've also seen ppl say quality is negatively affected if RVC is installed and ran from a drive other than the C drive. I don't understand how any of these aspects can affect quality but hopefully someone has insight?
azure marshBOT
RVC Guides (How to Make AI Cover)
thin stump
Does anyone know how hardware affects model training quality? I understand the d...
will quality be affected if training is done on a bad GPU vs good GPU
You can't use specific batch sizes. So if you have vram of 4, you can't use a batch size of 8
Some people only used batch size of 2's because that's all their pc can handle
somber gale
help
when i try to train my model and first epoch is about to finish, i get this error:
INFO:sxak:loss_disc=4.309, loss_gen=3.247, loss_fm=11.832,loss_mel=20.609, loss_kl=2.115
Process Process-1:
Traceback (most recent call last):
File "multiprocessing\process.py", line 315, in _bootstrap
File "multiprocessing\process.py", line 108, in run
File "C:\RVC\Mangio-RVC-v23.7.0\train_nsf_sim_cache_sid_load_pretrain.py", line 225, in run
train_and_evaluate(
File "C:\RVC\Mangio-RVC-v23.7.0\train_nsf_sim_cache_sid_load_pretrain.py", line 577, in train_and_evaluate
with open('csvdb/stop.csv') as CSVStop:
FileNotFoundError: [Errno 2] No such file or directory: 'csvdb/stop.csv'
someone please?
@thin stump
thin stump
someone please?
#✨│ai-help message Is this same error you have?
somber gale
https://discord.com/channels/1159260121998827560/1159290139609137264/12153480113...
nope
mine says "csvdb/stop.csv" no idea what that is its not in the directory
thin stump
ok, i'm just looking through the chat logs to be honest. You don't have a directory folder name csvdb?
somber gale
no, can u check if u do?
thin stump
no, can u check if u do?
I don't use local for training so yeah can't help
somber gale
dude i just want to make a model and i run into errors which nobody ever stumbled upon two times in a row wtf
what is csvdb/stop.csv???
why is the code trying to find that if its not in the directory
somber gale
I don't use local for training so yeah can't help
how do i train a model without mangio? mangio only makes problems kinda
thin stump
#✨│ai-help message Did you set up CSVDB?
thin stump
I mean, was there an option. yes
somber gale
the guy in my tutorial didnt mention anything
also im not using colab
when i search "setup csvdb" i only get shit results
rare gobletBOT
Ayo? @somber gale level 3 !!!
somber gale
@thin stump how do i setup csvdb?
thin stump
i don't know, sorry
brittle wing
Does anyone know how hardware affects model training quality? I understand the d...
for anyone wondering, i have found that hardware does matter in training quality. Lesser hardware that struggles with training will produce lesser quality models. If you do have capable hardware, quality will suffer if you are using the computer while training, ie., youtube, word, spotify, etc. my best models have come from leaving the computer running without using it for anything else, and applying settings that don't put too much strain on cpu/gpu. I haven't done testing specifically on external drives, but i also believe running rvc from C drive (ssd) as opposed to external hdd will improve performance.
finite galleon
why is the code trying to find that if its not in the directory
Try create a blank csv file named "stop.csv" that is placed in the csvdb folder
somber gale
Try create a blank csv file named "stop.csv" that is placed in the csvdb folder
that was my idea as well ill try
somber gale
Try create a blank csv file named "stop.csv" that is placed in the csvdb folder
still same error
finite galleon
still same error
Like the exact same file not found?
somber gale
ive done some digging, apparently nobody ever encountered this? why am i so unlucky?
somber gale
Like the exact same file not found?
yep
INFO:torch.nn.parallel.distributed:Reducer buckets have been rebuilt in this iteration.
INFO:sxak:Train Epoch: 1 [98%]
INFO:sxak:[54, 0.0001]
INFO:sxak:loss_disc=4.291, loss_gen=3.444, loss_fm=14.548,loss_mel=20.799, loss_kl=2.197
Process Process-1:
Traceback (most recent call last):
File "multiprocessing\process.py", line 315, in _bootstrap
File "multiprocessing\process.py", line 108, in run
File "C:\RVC\Mangio-RVC-v23.7.0\train_nsf_sim_cache_sid_load_pretrain.py", line 225, in run
train_and_evaluate(
File "C:\RVC\Mangio-RVC-v23.7.0\train_nsf_sim_cache_sid_load_pretrain.py", line 576, in train_and_evaluate
with open("csvdb/stop.csv") as CSVStop:
FileNotFoundError: [Errno 2] No such file or directory: 'csvdb/stop.csv'
On the RVC disconnected folder which uses the mangio fork, the setup csvdb is just creating a folder named "csvdb" and in it a "stop.csv" file (csvdb/stop.csv). The stop.csv contents is just False
somber gale
okay, i used my python skills and edited train_nsf_sim_cache_sid_load_pretrain.py:
I added,
if not os.path.isdir("csvdb"):
os.mkdir("csvdb")
with open("csvdb/stop.csv", "w+", newline="") as STOPCSVwrite:
csv_writer = csv.writer(STOPCSVwrite, delimiter=",")
csv_writer.writerow(["False"])
Did it work
somber gale
Im gonna try now
rugged stirrup
Anyone know how to get level audio that isnt like this?
somber gale
On the RVC disconnected folder which uses the mangio fork, the setup csvdb is ju...
there is nos setup csvdb thats the thinng
ill run it rq
finite galleon
there is nos setup csvdb thats the thinng
Yeah its a colab thing
somber gale
frick colab im gonna get this to work
in a second
nope didnt work
@finite galleon
finite galleon
<@818050831034613771>
You on windows or linux
finite galleon
win
Change the code to with open("csvdb\stop.csv") as CSVStop:
On linux paths uses / while windows uses \
somber gale
python paths use /
finite galleon
python paths use `/`
Pretty sure it depends on the os
somber gale
ill install an older versioooon of mangio which doesnt use csv this is bullshit
somber gale
what else do i use
finite galleon
Use Applio or Ilaria RVC
finite galleon
Or the mainline RVC
somber gale
or paid
somber gale
which do i use?
finite galleon
Any works
Some people like applio better than mainline
Sone dont
Its a matter of prefrence
somber gale
ill try applio
low shard
i just watched a random ahh youtube tutorial
99.999% yt tuts are outdated
Its better if u follow our docs instead
somber gale
Its better if u follow our docs instead
didnt even know yall had a fucking docs thank u man
low shard
Don't promo out of #1159290752195633273
low shard
didnt even know yall had a fucking docs thank u man
Yw
somber gale
Use Applio or Ilaria RVC
W, applio seems betteeer already its also faster and actually is managing to generate epochs currrently at 2/700
solar token
dusty wagon
hey guys, new to making ai covers, what do i do if i need a huggingface/google drive link but only have a applio download link?
rugged stirrup
You just need a Google account if you want to use colab If you want to run locally you dont need any account
zinc anchor
hey guys, new to making ai covers, what do i do if i need a huggingface/google d...
Do you have an option to upload a model from your PC?
In that case it's fine because the Applio bot usually makes you download the model on your PC
But in that page you can also see the original Huggingface link
zinc anchor
But in that page you can also see the original Huggingface link
The annoying thing is that you can't go over it with the mouse to copy and paste it later, you have to write it down
zinc anchor
When you have a finished model you can use it for inference
loud summit
Idk how to find a correct batch size
proper shale
Idk how to find a correct batch size
Just use 8 if you're using anything bigger than 10 minutes, 4 in case it's smaller
loud summit
Ok
proper shale
Also, it's always good to try different batch sizes when you can. Don't get stuck using only those, you can try out other numbers, like 6 and 12... just don't use odd numbers haha
That's the best way you can find the best batch for each model - experiment. Every model is different
azure marshBOT
RVC Guides (How to Make AI Cover)
heavy marten
scarlet rover
anybody else struggling with voice being choppy when broadcasted from client instead of server on discord?
foggy folio
-help
tall marten
Hi I'm currently working on covering anime songs with ai through rvc and AICoverGen
I am proceeding by separating the mp3 file into a main voice file and an instrumental file, applying the main voice file to the model, and combining it with the instrumental file
There is no problem with the voice until the point of separation by main voice, but as soon as the voice is applied to the model through AICoverGen, faint reverberation and mechanical sound are added to the voice
How can I resolve this problem?
grim nova
hey, i need help please i get this:
JSONDecodeError: Expecting value: line 1 column 1 (char 0)
low shard
hey, i need help please i get this:
JSONDecodeError: Expecting value: line 1 co...
ur prolly using an outdated colab, please use the updated ones inside the cloud folder inside the rvc folder of our docs https://docs.aihub.wtf
proper shale
Hi I'm currently working on covering anime songs with ai through rvc and AICover...
The faint reverb is applied in AICG outputs. Just change the reverb settings ig
Abt the mechanical sound, that can occur with certain models, ig you can use an audio editor to make that less apparent, and in the final mix it might be near invisible to the ears
grim nova
i keep getting the error on the step 2
JSONDecodeError Traceback (most recent call last)
<ipython-input-7-51a0b006a49c> in <cell line: 31>()
31 if os.path.exists(config_path):
32 # File exists, proceed with creation of creds and client
JSONDecodeError: Expecting value: line 1 column 1 (char 0)
proper shale
JSONDecodeError: Expecting value: line 1 column 1 (char 0)
might be using an outdated colab
easyGUI v2 (with crepe)?
grim nova
easyGUI v2 (with crepe)?
yeah i'm using that version
rare gobletBOT
Ayo? @grim nova level 1 !!!
proper shale
yeah i'm using that version
yea that's outdated and broken, check https://docs.aihub.wtf/
there's not only tutorials but colab links too
weak crypt
yea that's outdated and broken, check https://docs.aihub.wtf/
can you please link the latest easygui for me? i cant find that in that docs
weak crypt
thank you so much
proper shale
yw :)
small sorrel
we need to train the voices added by us right?
proper shale
yeah
small sorrel
we don't use voice.ai here
then
small sorrel
it supports RVC voices
proper shale
small sorrel
and is it safe?
proper shale
and is it safe?
yeah and yeah
in case u wanna check the code just search for w-okada on github, their repo seems pretty safe and I've used it a couple of times so yeh
small sorrel
in case u wanna check the code just search for w-okada on github, their repo see...
ohh okk
proper shale
check https://rentry.co/VoiceChangerGuide if you want the guide on installation and how to use it :)
small sorrel
check https://rentry.co/VoiceChangerGuide if you want the guide on installation ...
thnx
small sorrel
do keep in mind that you need an ok-ish GPU to run
rtx 2060 siuper
proper shale
oh yeah you should be fine then!
small sorrel
small sorrel
oh yeah you should be fine then!
alright
serene hound
is there a way to get more fluent language tts bases for applio?
small sorrel
oh yeah you should be fine then!
okada is a W man
rare gobletBOT
Ayo? @small sorrel level 2 !!!
dreamy bramble
Hey, i have this error all the time when creating a model: NameError Traceback (most recent call last)
<ipython-input-12-ffaaf2c2b4e7> in <cell line: 27>()
25
26 pretrained_base = "pretrained/" if model_architecture == "v1" else "pretrained_v2/"
---> 27 unpt = f"_{pretrain_type}" if pretrain_type!="original" else ""
28
29 pretrainedD = f"{pretrained_base}f0D{target_sample_rate}{unpt}.pth"
NameError: name 'pretrain_type' is not defined
Can somone tell me how to solve it?
odd shale
Hey, i have this error all the time when creating a model: NameError ...
Show your rvc window.
If it says "pretrain_type" maybe the path file to the pretrain that should be used for training your model is empty
crimson basalt
Traceback (most recent call last):
File "E:\Ilaria RVC\runtime\lib\site-packages\gradio\routes.py", line 321, in run_predict
output = await app.blocks.process_api(
File "E:\Ilaria RVC\runtime\lib\site-packages\gradio\blocks.py", line 1007, in process_api
data = self.postprocess_data(fn_index, result["prediction"], state)
File "E:\Ilaria RVC\runtime\lib\site-packages\gradio\blocks.py", line 953, in postprocess_data
prediction_value = block.postprocess(prediction_value)
File "E:\Ilaria RVC\runtime\lib\site-packages\gradio\components.py", line 2076, in postprocess
processing_utils.audio_to_file(sample_rate, data, file.name)
File "E:\Ilaria RVC\runtime\lib\site-packages\gradio\processing_utils.py", line 206, in audio_to_file
data = convert_to_16_bit_wav(data)
File "E:\Ilaria RVC\runtime\lib\site-packages\gradio\processing_utils.py", line 219, in convert_to_16_bit_wav
if data.dtype in [np.float64, np.float32, np.float16]:
AttributeError: 'NoneType' object has no attribute 'dtype'
ilaria rvc worked once now im getting this?
boreal shore
I have problem. Can somebody help?
odd shale
ilaria rvc worked once now im getting this?
Ig Ilaria RVC Colab is broken again
crimson basalt
Ig Ilaria RVC Colab is broken again
im using local
odd shale
im using local
Show your RVC GUI anyway
crimson basalt
Show your RVC GUI anyway
hmm. i dont get this error if i use the upload audio option. but picking audio from the drop down gives me an error
rigid jewel
for anyone wondering, i have found that hardware does matter in training quality...
I would think the opposite... Having to train at a lower batch size would make your models higher quality no?
odd shale
hmm. i dont get this error if i use the upload audio option. but picking audio f...
It's a weird error then.
proper shale
I would think the opposite... Having to train at a lower batch size would make y...
Not necessarily, it kinda makes the training more unstable if you go too low. You have to find the best batch for each model, and this involves a lot of testing and messing around ig
proper shale
Hey, i have this error all the time when creating a model: NameError ...
Did you run "Set Training Variables"?
craggy light
Help me, I can’t get the program to work with me. I’ve tried everything and it still won’t accept any model I’m trying to use
rare gobletBOT
Ayo? @craggy light level 2 !!!
odd shale
Did you run "Set Training Variables"?
🐢
proper shale
Help me, I can’t get the program to work with me. I’ve tried everything and it s...
what program are you using specifically?
craggy light
I’m trying to use the collab link for RVC
proper shale
I’m trying to use the collab link for RVC
ya might be using an old one
is it the easy GUI v2 (with crepe)?
craggy light
Yes
proper shale
yeah that's old and has been outdated for a while
craggy light
Ah I see, that explains a lot
proper shale
https://docs.aihub.wtf/ has colab links + tutorials, check out that page whenever you need links :)
bookmark it
or something
craggy light
Thanks
proper shale
You're welcome :)
azure marshBOT
RVC Guides (How to Make AI Cover)
grim nova
omg i tried using ilaria, rvc local... still can't make it work
i think it's a model problem
i don't know
grim nova
this is so frustrating
proven hill
whats the problem
grim nova
whats the problem
i can't convert in rvc local, it says error and i followed all the steps idk what i did wrong
this is the model i've been trying to use https://huggingface.co/MusologyWavey/MusologyRVC/resolve/main/Dan Reynolds.zip?download=true
wispy burrow
i can't convert in rvc local, it says error and i followed all the steps idk wha...
What is the error that's displayed in the cmd?
grim nova
What is the error that's displayed in the cmd?
AttributeError: 'NoneType' object has no attribute 'dtype'
wispy burrow
AttributeError: 'NoneType' object has no attribute 'dtype'
I usually see that error here when the path to the audio file is wrong or something like that.
proven hill
i can't convert in rvc local, it says error and i followed all the steps idk wha...
rvc local as in the classic one or the ilaria rvc mainline?
grim nova
rvc local as in the classic one or the ilaria rvc mainline?
rvc local
grim nova
She's asking which version you use locally.
Regular RVC or Ilaria-RVC.
oh lmao, sorry for being dumb on this, it says RVC WebUI
rare gobletBOT
Ayo? @grim nova level 2 !!!
grim nova
oh lmao, sorry for being dumb on this, it says RVC WebUI
so i think it's regular rvc
rich marsh
Hey, i have this error all the time when creating a model: NameError ...
you skip "Download Pretrained Model" . thats why
rare gobletBOT
Ayo? @rich marsh level 1 !!!
wispy burrow
so i think it's regular rvc
Can you upload a screenshot of RVC?
proper shale
yep thats the latest one
midnight crest
quick question on applio, is there a known text limit for .txt files uploaded to tts? i tried loading a small ebook and it keeps failing, but was able to get it to do 5 minutes of word content at a time
tried checking docs and github and couldnt see any notes on it
proper shale
quick question on applio, is there a known text limit for .txt files uploaded to...
i think there might be a limit
or maybe a hardware limitation? hmm
nova trench
Hello, terribly sorry if I'm in the wrong channel, but can I ask questions related to so-vits-svc here? Or is there another dedicated channel for that? Sincerest apologies, but I'm kinda new to the whole Discord thing ;; I'm having an issue with voice infering
proper shale
not sure tbf
proper shale
Hello, terribly sorry if I'm in the wrong channel, but can I ask questions relat...
uhh
hate to say this but... I don't think there's any reason to keep using sovits
midnight crest
not sure tbf
no worries! thanks for your input
proper shale
no worries! thanks for your input
yw :)
compact sinew
uhh
hate to say this but... I don't think there's any reason to keep using sovi...
probs they using sovits 5.0
atomic mulch
yep thats the latest one
thanks!
proper shale
probs they using sovits 5.0
it's still... kinda outdated at this point
nova trench
uhh
hate to say this but... I don't think there's any reason to keep using sovi...
Oh wait? Can I ask why? I've one of my favorite cover-makers utilize 4.1, and I really wanted to try it for myself orz
rare gobletBOT
Ayo? @nova trench level 1 !!!
compact sinew
it's still... kinda outdated at this point
yeah fax
proper shale
but to be honest rvc v2 is almost a year old so
compact sinew
Oh wait? Can I ask why? I've one of my favorite cover-makers utilize 4.1, and I ...
takes forever to train and you need more data than rvc
and its not as good
proper shale
^
nova trench
takes forever to train and you need more data than rvc
Oooh, I see I see... I'll try to utilize RVC, then o7 Should I uninstall so-vits, then?
compact sinew
Oooh, I see I see... I'll try to utilize RVC, then o7 Should I uninstall so-vits...
if u want to
nova trench
Because all I'm really getting is an error about being unable to import kaiser from scipy.signal
brittle wing
what is a dataset?
imagine we need a tutorial video on how do make an ai model on Illaria-RVC
quasi elbow
has anybody else had any problems with created models sounding TOO similar to the given audio file? I've noticed that a model I've made recently tends to copy the general intonation and tone of the original singer way more than a model of the same person I found on the internet about a month ago. i havent really looked at RVC developments much in the past few months though so i'm just training like how I usually do - not sure if any new advancements have come out that make models sound more like "themselves" so to speak.
my model uses a ~30 minute dataset at 200 epochs if it helps
proper shale
what is a dataset?
the data that you use to train an ai model, so in this case audio files
proper shale
has anybody else had any problems with created models sounding TOO similar to th...
depending on the type of voice it might not do a lot of changes
quasi elbow
depending on the type of voice it might not do a lot of changes
do you know what would cause two different models of the same voice to have differing "strength" in terms of similarity to the original? the online model seems to have been made from a more poorly isolated/LQ dataset but stays more true to the singer its modelled after
would dataset volume have any impact maybe?
proper shale
would dataset volume have any impact maybe?
most likely the dataset's contents... also, experiment with batch_size 4, sometimes the model wasn't cooked properly?
quasi elbow
most likely the dataset's contents... also, experiment with batch_size 4, someti...
ill train another model tonight with batch_size 4...also following the tensorboard during training, my model seemed to be best optimized very early on (~50 epochs in). would that be cause for concern?
rare gobletBOT
Ayo? @quasi elbow level 1 !!!
proper shale
ill train another model tonight with batch_size 4...also following the tensorboa...
Nah not really, it's still learning so idk
midnight crest
i think there might be a limit
following up rn but it looks like the limit for applio tts .txt files is somewhere around 30,608 characters with spacing, about 5,389 words
able to produce about 32:30 minutes of tts audio output at a time. about 10 pages in word doc
proper shale
damn
midnight crest
ill see if i can add that note in for the applio documentation or something, figure it may be helpful for folks to know
rare gobletBOT
Ayo? @midnight crest level 1 !!!
tawny wigeon
Does anyone have a Google Colab link so I can train my own model?
proper shale
Here's RVC Disconnected:
https://colab.research.google.com/drive/1XIPCP9ken63S7M6b5ui1b36Cs17sP-NS
Definitely check out the tutorial for it:
And btw, a little tip for you:
Limit the training session to 1 or 2 hours... if you train for too long you get disconnected and lose your model
scenic arch
tawny wigeon
Here's RVC Disconnected:
https://colab.research.google.com/drive/1XIPCP9ken63S7...
thanks!!!!!
proper shale
thanks!!!!!
yw!
velvet finch
why i cant get a cs2 agent vioce model
elder estuary
why i cant get a cs2 agent vioce model
i dont think anyone made one yet
velvet finch
oh oke
rare gobletBOT
Ayo? @velvet finch level 1 !!!
rustic locust
i did
thin stump
Want to try again? Reload your disconnected colab session
and run the dependecies again
rustic locust
it says cannot find file
thin stump
Make sure your zip folder name has no spaces or special characters when you upload that on google drive
rustic locust
okay
rare gobletBOT
Ayo? @rustic locust level 2 !!!
rustic locust
i got it to work
rare gobletBOT
Ayo? @tawny wigeon level 1 !!!
rustic locust
now its doing the processing and feature extraction
thin stump
how do i save this model for later use? I can't
You only get 4 hours until you get kicked out so even if you save preproccessing files, you wouldn't have a checkpoint
assuming this is for colab
rustic locust
IT'S WORKING
thin stump
Remember to train the index
rustic locust
I will
proper shale
how do i save this model for later use? I can't
stop training and export to google drive, or even better Huggingface
thin stump
ok, here's the tensorboard page https://docs.aihub.wtf/rvc/resources/epochs--tensorboard/#-setting-up
You're only monitoring for g/total. That's it for newbies
thin stump
do i just let that run for a while?
Yeah, it'll take a while. I think captchas showing up will vary for different people
It wouldn't matter if you click on anything, google will still ask you if you're there
stable crater
something wrong with the collab on 2nd step
rustic locust
wut
thin stump
do i just let that run for a while?
but just remember you only get 4 hours of colab gpu
thin stump
so you have to stop the training before you run out, ye
then all your files get saved in
Mangio Fork
Weights
rustic locust
so then how do I get them?
thin stump
@proper shale I don't know much about that since I never resumed a training session before
lavish minnow
How can you create a model
rustic locust
my g/total graph is not updating
undone fractal
hello guys im new here, I installed easy GUI but would like to swap my voice with an artist one. How to do that ? On voice models section, i have link but dont know how to use it on google collab. Thanks in advance
brittle wing
how can i have the montanablack88 voice?
thin stump
How can you create a model
Prepare a dataset first. https://docs.aihub.wtf/essentials/how-to-make-voice-models/ @brittle wing make it yourself
thin stump
my g/total graph is not updating
You didn't turn on the 30 seconds cog
rustic locust
and the percentage is going down
brittle wing
what
thin stump
You didn't turn on the 30 seconds cog
brittle wing
yeah i found it
rare gobletBOT
Ayo? @brittle wing level 1 !!!
brittle wing
🖕
rare gobletBOT
Ayo? @lavish minnow level 1 !!!
stable crater
the colab is just not working when i try to download the voice model
fails instantly
stable crater
is this a fucking circus? actually help me!
lavish minnow
Monte Diggi
brittle wing
thin stump
is this a fucking circus? actually help me!
and no I won't be able to help until you get your image perms at level 1
