i dont think 3050 is enough
#✨│ai-help
molten pecan
Ask in #🔍│help-w-okada
molten pecan
i dont think 3050 is enough
Wait what really?
blissful basin
Ask in <#1159290161683767298>
I already got the okada one i want the other because my mic keep cutting out on this one
molten pecan
I already got the okada one i want the other because my mic keep cutting out on ...
The okada channel is for real-time inference in general
blissful basin
so what is this one for?
rare gobletBOT
Ayo? @blissful basin level 1 !!! 
molten pecan
so what is this one for?
Training models/inference/setting up RVC
strange quest
blissful basin
oh
molten pecan
@dawn nebula what do you say?
hot lagoon
I try to follow this but the pich at some part gose hight
molten pecan
😭
strange quest
strange quest
thats not a 3050
now im confused
blissful basin
could storage affect the voice changer?
strange quest
am i still ok lol
dawn nebula
am i still ok lol
yeah ofc keep training
strange quest
thank you ❤️
the real question is... how do i stop 
i think the tut said this would change to stop but it didnt
im assuming cuz its saving automatically i can just close when im ready?
dawn nebula
im assuming cuz its saving automatically i can just close when im ready?
send me your training settings
molten pecan
i think the tut said this would change to stop but it didnt
The tut says to simply close RVC if that were to happen
strange quest
thank youu
its only on 78 now so im just leaving it to it and checking every so often
like 1 epoch a minute
solemn dagger
Does batch size affect training speed?
pallid hare
How do I make the models not sound like... This?
(It's eating the words, it sounds extremely nasal even)
thin stump
How do I make the models not sound like... This?
Did you denoise your dataset with MVSEP? Using MDX23 then Melband reformer
thin stump
https://tenor.com/view/huh-cat-huh-m4rtin-huh-huh-meme-what-cat-gif-27377993
That's why. RVC is training on the noises and not the voice https://mvsep.com/en and read https://aihubdocs.github.io/en/vocal-isolation--datasets/datasets/#datasets
pallid hare
How do I make the models not sound like... This?
Also uh
That's not an acapella from an existing song, it's an original song
wind urchin
hey,
why is rvcDisconnected not training features on my dataset? I run preprocessing and feature extraction and get this
https://cdn.aayus.me/uploads/firefox_49ng9rwO7H.png
and when i create a .index, i get this
https://cdn.aayus.me/uploads/firefox_ypavIMsi8U.png
thin stump
Does batch size affect training speed?
It does. I would stick to 8 unless you have a bigger dataset so you can justify using 12
pallid hare
eh I'm just finding another model
thin stump
eh I'm just finding another model
Ohh, I thought you were training a model
You usually can't fix it except change the index or fetch ratio down to 0.4
thin stump
hey,
why is rvcDisconnected not training features on my dataset? I run preproces...
refresh and redo the dependecies?
wind urchin
refresh and redo the dependecies?
done, ive even deleted the dataset .zip and reuploaded. tried a different name
obtuse stream
anyone help me what thease error?
azure marshBOT
RVC Guides (How to Make AI Cover)
solemn dagger
I'm trying to train a model I've train previously, but when I go to train it starts generating the epochs from the beginning. The Epochs will start at 2 instead of where it left off. I'm using Applio. Is this normal?
Do I need to preprocess, extract and generate the index again? (Edit: I'm just going to start over lol)
wise estuary
I downloaded a video from Youtube to train a model with the voice, the format was mp4 and i converted it to .wav
Is that useless or do i get more than mp3?
low shard
I downloaded a video from Youtube to train a model with the voice, the format wa...
you need only to clean the voice so theres no noises or instrumentals
its better you use .wav
not mp3
glossy grove
so im trying to use the model fusion on applio 2. i keep getting this error after pointing it to the two models im attepting to fuse Traceback (most recent call last): File "/home/fletcher/Music/Applio-RVC-Fork_Tweaks/lib/infer/infer_libs/train/process_ckpt.py", line 220, in merge cfg = ckpt1["config"] KeyError: 'config'
placid rivet
Can you suggest project TTS use RVC Model ?
glossy grove
still not working on 3.0.8 ```Traceback (most recent call last):
File "/home/fletcher/Music/Applio-3.0.8/.venv/lib/python3.10/site-packages/gradio/queueing.py", line 495, in call_prediction
output = await route_utils.call_process_api(
File "/home/fletcher/Music/Applio-3.0.8/.venv/lib/python3.10/site-packages/gradio/route_utils.py", line 232, in call_process_api
output = await app.get_blocks().process_api(
File "/home/fletcher/Music/Applio-3.0.8/.venv/lib/python3.10/site-packages/gradio/blocks.py", line 1561, in process_api
result = await self.call_function(
File "/home/fletcher/Music/Applio-3.0.8/.venv/lib/python3.10/site-packages/gradio/blocks.py", line 1179, in call_function
prediction = await anyio.to_thread.run_sync(
File "/home/fletcher/Music/Applio-3.0.8/.venv/lib/python3.10/site-packages/anyio/to_thread.py", line 56, in run_sync
return await get_async_backend().run_sync_in_worker_thread(
File "/home/fletcher/Music/Applio-3.0.8/.venv/lib/python3.10/site-packages/anyio/_backends/_asyncio.py", line 2144, in run_sync_in_worker_thread
return await future
File "/home/fletcher/Music/Applio-3.0.8/.venv/lib/python3.10/site-packages/anyio/_backends/_asyncio.py", line 851, in run
result = context.run(func, *args)
File "/home/fletcher/Music/Applio-3.0.8/.venv/lib/python3.10/site-packages/gradio/utils.py", line 678, in wrapper
response = f(*args, **kwargs)
File "/home/fletcher/Music/Applio-3.0.8/core.py", line 397, in run_model_blender_script
message, model_blended = model_blender(model_name, pth_path_1, pth_path_2, ratio)
TypeError: cannot unpack non-iterable KeyError object
there is a tts mixed with rv
which is edge tts at the moment
check #1212368971127590922
copper marlin
hi
rare gobletBOT
Ayo? @copper marlin level 10 !!!
copper marlin
how do i transfer voices made in one installation to another?
I tried transfering the named folders from the logs but that didnt work
only the voices trained with the installation were showing up
couldnt find the .pth file either
assets/weights
probs
forest sluice
Does batch size affect training speed?
yes
solemn dagger
Thank you Ouchy
onyx ferry
°
thick dagger
-rvc
marble ruin
I have a question
My voice changer isn't working correctly, whenever I click that start.bat, it closed automatically.
Nothing else popped up except the commands and closed itself.
neon mortar
question if i wanna update or change the ver of rvc do i just extract to the same file or i have to delete first
magic frigate
hi, are there male voices? because I didn't find it
latent kettle
hi, are there male voices? because I didn't find it
search on weights.gg
azure marshBOT
RVC Guides (How to Make AI Cover)
magic frigate
search on weights.gg
thx u
low shard
My voice changer isn't working correctly, whenever I click that start.bat, it cl...
ask wokada questions in #🔍│help-w-okada
brittle wing
Hey, question since I'm not familiar with how GPUs work and such. If I connect a nvidia eGPU to my AMD laptop, will I be able to run RVC locally?
proper shale
Hey, question since I'm not familiar with how GPUs work and such. If I connect a...
really good question
honestly, idk if anybody has ever tried that
it might work, but I can't guarantee anything
obsidian dew
Traceback (most recent call last):
File "C:\Program Files (x86)\RVC-GUI-pkg\rvcgui.py", line 129, in vc_single
audio_opt = vc.pipeline(
File "C:\Program Files (x86)\RVC-GUI-pkg\vc_infer_pipeline.py", line 326, in pipeline
sid = torch.tensor(sid, device=self.device).unsqueeze(0).long0)
RuntimeError: CUDA error: the launch timed out and was terminated
CUDA kernel errors might be asynchronously reported at some other API 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.
Hello. Ireceived this error message on rvc desktop app. How can I solve this?
brittle wing
really good question
honestly, idk if anybody has ever tried that
interesting. i'll use my friend's eGPU and report back my results
vocal hedge
I do AI covers through IlariaRVC(huggingface). An error. What to do?
vocal hedge
-colab
molten pecan
I do AI covers through IlariaRVC(huggingface). An error. What to do?
the HF space is kinda obsolete now. use Colab. it's much much faster
orrrr the Applio Colab
=3
green citrus
how to setup tensorboard for rvc local? is this possible?
merry sandal
Guys, hello. The question is, if I want to deploy AiCoverGen on a server, how much power do I need in order for 5 people to convert a 3 min track at the same time?
rare gobletBOT
Ayo? @merry sandal level 1 !!!
left oasis
are there any other sites for creating ai besides hina?
left oasis
https://aihubdocs.github.io/en/rvc/cloud/ilaria-rvc/
it's work?
dull furnace
may i know why I use voice changer client demo and upload a pth but nothing work?
the option doesn't pop up
mighty flame
When I'm on the Google Colab, and i reach step 2, I get the error:
JSONDecodeError Traceback (most recent call last)
<ipython-input-16-348ecd18811f> 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)
idk what this mean
viscid moss
When I'm on the Google Colab, and i reach step 2, I get the error:
JSONDecodeEr...
Hi! U re using an outdated colab, use the new ones
-colab @mighty flame
azure marshBOT
-colab <@621688056562319381>
Suggestions for @mighty flame
viscid moss
Guides here:
https://aihubdocs.github.io/en
mighty flame
Guides here:
https://aihubdocs.github.io/en
Thanks!
Which one do I use for making an AI Cover?
foggy folio
why isn't this pinned
viscid moss
Thanks!
Which one do I use for making an AI Cover?
For cover u can use:
Ilaria RVC or Applio
If u re too lazy and u don't want to separate the instrumental, reverb, etc. U can use AI Covergen
wind cedar
someone can help me? when i try starting rvc, it says, vcc client - wait web server
thin stump
Eh it's ok. It's what worked for me
mighty flame
For cover u can use:
Ilaria RVC or Applio
If u re too lazy and u don't want to s...
Ty :D
rare gobletBOT
Ayo? @mighty flame level 1 !!!
thin stump
I pulled all the information from this thread and condensed it https://discord.com/channels/1159260121998827560/1201368092358492210
viscid moss
someone can help me? when i try starting rvc, it says, vcc client - wait web ser...
try deleting the “stored_settings.json” file in the “MMVCServerSIO” folder and restart the app
foggy folio
guys is there a way to skip this shitty mvsep queue without buying premium
site used to be faster and no premium
foggy folio
melband reform or something
meager stirrup
I need to make cover tweedledum part in twisted twins fnf song but it's too much high pitched can you tell me how to fix a pitch of singer to make it normal?
viscid moss
I have made UVR5 colabs, but melband doesn't works <:AIHC_Ded:116380751689600211...
If u wanna test
Demucs:
https://colab.research.google.com/github/Eddycrack864/Demucs-for-Colab/blob/main/Demucs_for_Google_Colab.ipynb
UVR5 Models (even VIP models)
https://colab.research.google.com/github/Eddycrack864/Ultimate-Vocal-Remover-5.6-for-Google-Colab/blob/main/Ultimate_Vocal_Remover_5_6_for_Google_Colab.ipynb
Another models (MDX23C, Vit Large, etc)
https://colab.research.google.com/github/Eddycrack864/Music-Source-Separation-Universal-Colab/blob/main/Music_Source_Separation_Universal_Colab.ipynb
viscid moss
I need to make cover tweedledum part in twisted twins fnf song but it's too much...
U can change the pitch
viscid moss
0 More high pitch
<0 More low pitch
thin stump
Also I said BVE 2nd since melband might strip away backings as a noise. It's all dependent on the song
mighty flame
For cover u can use:
Ilaria RVC or Applio
If u re too lazy and u don't want to s...
I'm struggling with how to insert multiple files for either :(
thorn tree
I followed the guide and ended up on this discord. I got something working but its only cut off fragments and doesnt fully do sentences :/
2024-03-06 18:16:48 | INFO | faiss.loader | Successfully loaded faiss with AVX2 support.
2024-03-06 18:16:48 | INFO | configs.config | Found GPU NVIDIA GeForce RTX 4070
is_half:True, device:cuda:0
Input device: 2:Microphone (2- JBL Quantum 810 (MME)
Output device: 6:CABLE Input (VB-Audio Virtual C (MME)
cuda_is_available: True
Index search FAILED or disabled
Loading rmvpe model
Audio block passed.
Audio block passed.
Spent time: fea = 0.15s, index = 0.00s, f0 = 0.92s, model = 0.64s
sola_offset = 0
Infer time: 1.83
Audio block passed.```
is this console output correct?They convert one by one 
mighty flame
oh
mighty flame
They convert one by one <:nails:1159569314848972891>
ty
viscid moss
still not found a good way to get rid of backing vocals
Did you try with BVE or 5/6 HP?
foggy folio
Did you try with BVE or 5/6 HP?
bve removes all the vocals
I haven't tried 5hp
but 6 keeps all of the backing vocals
mighty flame
is this normal? I've done everything like the guide said (currently converting)
viscid moss
is this normal? I've done everything like the guide said (currently converting)
What colab or space are u using?
verbal basin
is ilaria down?
rare gobletBOT
Ayo? @verbal basin level 4 !!!
mighty flame
I've added the voice model as intended and added the model
viscid moss
https://huggingface.co/spaces/TheStinger/Ilaria_RVC
yep, it's normal. HF Spaces uses CPU only
mighty flame
its prob because i dont have an acc 🤦♂️
viscid moss
No GPU
mighty flame
yep, it's normal. HF Spaces uses CPU only
ah ok
viscid moss
its prob because i dont have an acc 🤦♂️
also this
viscid moss
Instead of HF Spaces u can use Colab
mighty flame
which colab should i use?
molten pecan
which colab should i use?
https://aihubdocs.github.io/en/ the Applio Colab or Ilaria RVC ones, explained here =3
mighty flame
thanks :D
rare gobletBOT
Ayo? @mighty flame level 2 !!!
mighty flame
yw ^-^ have fun
trying both methods, the hf space and the new one :D
molten pecan
trying both methods, the hf space and the new one :D
The HF space is much much slower compared with Colab. It's kind of obsolete
So I recommend sticking with Colab when you can
thorn tree
So uh no clue then? As im running locally on a 4070 i kinda hoped this wouldve worked :/
mighty flame
ok friendly reminder for me in the future, DO NOT use eminem with a high-pitched voice 💀
azure marshBOT
RVC Guides (How to Make AI Cover)
foggy folio
"use bve model for extracting main vocals!" bve:
brittle wing
How tf do i get the hugging face link of voice models
viscid moss
How tf do i get the hugging face link of voice models
-upload
viscid moss
damn images are broken 
thorn tree
Getting voice cutoff at the end of sentences but looks like it may be doing a thing
rare gobletBOT
Ayo? @thorn tree level 1 !!!
viscid moss
How tf do i get the hugging face link of voice models
Just right click on the download icon and then copy the link
brittle wing
So basically do you download models from weights.gg, upload them to huggingface then copy the link from there?
foggy folio
guys what settings should I use for bve to actually work
like window size and aggression
tf
this server is dead but kinda not dead at the same time
2 minutes by now 🙂
foggy folio
guys what settings should I use for bve to actually work
this was the question btw
past parrot
yoo
foggy folio
guys what settings should I use for bve to actually work
it has been 11 minutes 😭
brittle wing
So basically do you download models from weights.gg, upload them to huggingface ...
Yeah I'm sorry but this system is kinda eh
What happened to the old one where ppl who make models jst include the huggingface link aswell? 💀
viscid moss
What happened to the old one where ppl who make models jst include the huggingfa...
u can found the HF links of the models on #1175430844685484042
viscid moss
Seems to be down atm, try AICovergen or Applio
foggy folio
17 MINUTES NO ANSWER
edgy sky
is it normal that ilaria rvc on huggingface is taking more than 25 minutes to revoice a 3 minutes wav file?
viscid moss
like window size and aggression
Well, I always use BVE on MVSEP with 0,5 of aggression. Window size idk tbh
viscid moss
is it normal that ilaria rvc on huggingface is taking more than 25 minutes to re...
25? Did u duplicate the space?
edgy sky
25? Did u duplicate the space?
yep
foggy folio
Well, I always use BVE on MVSEP with 0,5 of aggression. Window size idk tbh
sadly I can't use that on local uvr
toxic ermine
is ilaria rvc broken?
it says Traceback (most recent call last): File "/content/Ilaria-RVC/Ilaria-RVC/app.py", line 1, in <module> import subprocess, torch, os, traceback, sys, warnings, shutil, numpy as np ModuleNotFoundError: No module named 'torch'
sand iris
to the top, there's a thing named "runtime"
under it "disconnect and delete runtime"
oh also, did you run the two cells on colab ?
Or did you go straight to launching it without doing anything before
half cove
Or did you go straight to launching it without doing anything before
would that not give app.py doesnt exist error or whatever
sand iris
nvm yes
tacit anvil
can you use datasets as the models in rvc?
half cove
can you use datasets as the models in rvc?
you use a dataset to create a model
tacit anvil
and is there a dataset to model converter?
foggy folio
and is there a dataset to model converter?
obviously
half cove
and is there a dataset to model converter?
by training a model using the dataset you get the model
edgy sky
25? Did u duplicate the space?
hey, what kind of link should i put in ilaria rvc to upload the model? i tried drive, mega and huggingface and none helps
left lantern
is covergen down ^^?
viscid moss
is covergen down ^^?
Nop
dapper hemlock
Anybody know where I can use tts with custom models?
tacit anvil
<https://aihubdocs.github.io/en/essentials/how-to-make-voice-models/>
so i go through this entire page until what step?
half cove
it explains
pale moat
viscid moss
hey, what kind of link should i put in ilaria rvc to upload the model? i tried d...
Hmmm restart the HF Space
foggy folio
where do I put voice models inside of applio local
ohh cool
ty for that
weights.gg covers don't hit the same
tacit anvil
it explains
what if i can't wait 2:30 hours?
half cove
thats just how long it takes
tacit anvil
is there a way to hurry it up?
foggy folio
is there a way to hurry it up?
no
rare gobletBOT
Ayo? @tacit anvil level 3 !!!
tacit anvil
i have an idea
how about you convert my datasets to rvc models
and send them into my dms
then i test them
tacit anvil
about 2000 epohs
tacit anvil
i'll send the datasets into your dms
foggy folio
sure!
obsidian dew
Traceback (most recent call last):
File "C:\Program Files (x86)\RVC-GUI-pkg\rvcg...
+1
fast swan
Rate an idiot's first time making a model ever
obtuse stream
anyone help me what thease error?
Anyone???
foggy folio
guys if I overtrain something will the backups from previous epochs save my ass?
foggy folio
(sent by PH Arthuro games 2024)
bronze lintel
Yo guys I’m just wondering if anyone knows how to help me with my problem
rare gobletBOT
Ayo? @bronze lintel level 1 !!!
bronze lintel
I followed the laria rvc doc instructions, but when i heard the “doofenshmirtz” voice it sounded really not great
ill send what it sounds like
foggy folio
lol my message got deleted
and my question never got answered
fuck mods
oh no mods and admins are gonna get heated at me now 😱
odd shale
oh no mods and admins are gonna get heated at me now 😱
I know you were joking but you shouldn't insult people.
foggy folio
I know you were joking but you shouldn't insult people.
aw sadly I am unhinged asf
I might have slipped out that word beginning with n in a few arguments
bronze lintel
I followed the laria rvc doc instructions, but when i heard the “doofenshmirtz” ...
Can anyone help with this?
trail prawn
I'm looking for emanuelle macron model
dawn nebula
I might have slipped out that word beginning with n in a few arguments
and you know the consequences
bronze lintel
ive tried changing the model but nothing changed
tawny granite
Expecting value: line 1 column 1 (char 0) error
rare gobletBOT
Ayo? @tawny granite level 1 !!!
foggy folio
and you know the consequences
"THERE ARE CONSEQUENCES FOR SAYING THE N WORD!!" "HOW DO YOU KNOW?? HOW DO YOU KNOOOOOOOOWWWWWW!!!!!!!!!!!"
rare gobletBOT
Ayo? @foggy folio level 6 !!!
azure marshBOT
All-In-One Guide on how to make a good model
This guide explains how the D and G files works and much more: https://rentry.org/RVC_making-models
Credits: LUSBERT 
Automated Overtraining Detection (AOD)
Will be available soon in #1159513888199540817
Credits: grvyscale
foggy folio
ah ty
I made a luigi model with someones dataset and??
it sounded like the terminator
having a stroke
thin stump
Should I stop? This looks like a really bad collapse
dawn nebula
Should I stop? This looks like a really bad collapse
zoom out
magic bronze
Hello, my voice sounds a bit distorted when I am using the voice changer. For context, I am new to this, have a nvidia 970m, have my tune to 12 and my index to 1
thin stump
zoom out
nvm I'll keep going as long as it goes down. I never had it this shaky in the initial slope before
azure marshBOT
RVC Guides (How to Make AI Cover)
glass jolt
After restarting PC to install VB audio cable, I re open the software and it’s just a white screen. What do I do?
hot lagoon
is there a way to do voice changer paperspace?
spare shale
is there a way to do voice changer paperspace?
??
odd shale
is there a way to do voice changer paperspace?
No, you can't
Paperspace forbids usage of ngrok tunnels.
If you use ngrok tunnels on Paperspace, you'll get banned
hot lagoon
is there a workaroud to it?
spring orchid
-colab
thin stump
is there a workaroud to it?
What's wrong with using hina colab?
wispy lion
if my dataset is 1min long or less how many times should i train it for best results?
stoic cloud
use headphones
rare gobletBOT
Ayo? @brittle wing level 1 !!!
stoic cloud
use ones that do not bleed sound
thin stump
if my dataset is 1min long or less how many times should i train it for best res...
Use the OV2 pretraining and look at the tensorboard. g/total will keep going down until it rises indefinetly
stoic cloud
if it didnt, it doesnt change your friends' voices in vc
Stop pinging me
If your headphones bleed sound, then get different ones or turn down the volume
livid gale
hi
low cliff
why won't google colab work?
rare gobletBOT
Ayo? @low cliff level 3 !!!
latent kettle
why won't google colab work?
which notebook are you using?
low cliff
which notebook are you using?
I'm using easy GUI and RVC Disconnected
latent kettle
I'm using easy GUI and RVC Disconnected
rvcDisconnect works. but not sure about Easygui
dusty bramble
question
of rvcdisconnect disconnects, is there any way to download the voice after or nah?
viscid moss
of rvcdisconnect disconnects, is there any way to download the voice after or na...
nah
whole lynx
does anyone have the code to make my google colab not disconnect?
wise estuary
I trained a model over the night, every time i looked it said suc
Now i stand up and go to my pc to see after the model and it said it’s finished but i dont have any index file and 2 path files wich booth not work with wokada
Is there anything i missed?
Update: The 2 path files i mentioned above are named G_xxxxxx and D_xxxxxx the x resembles number gibberish. Both files are around 350 and 600 kb big.
Now i looked through the files and found in the weight folder ava.path, (thats my model name) and it’s something around 60 kb big.
Does anyone know where i find the index file?
azure marshBOT
RVC Guides (How to Make AI Cover)
undone stratus
Is it compatible with Windows 11? because I can't install it
wise estuary
Is it compatible with Windows 11? because I can't install it
Im using Windows 11 and it worked so good how i mentioned a message above.
But it seems to do ot all fine
undone stratus
Im using Windows 11 and it worked so good how i mentioned a message above.
But i...
Which version are you using? please
wise estuary
Which version are you using? please
https://youtu.be/Q8du7n0vgfU?si=jwalqkg35QHtuthz I used this app from the tutorial
sand island
this is probably slightly unrelated but for those who uses a audio slicer to slice your audio files/datasets what audio slicer do you guys use? Ive been using the one by flutydeer without problem until recently where it just refuses to slice my audios and stuck at 0%
undone stratus
https://youtu.be/Q8du7n0vgfU?si=jwalqkg35QHtuthz I used this app from the tutori...
Thanks u
digital epoch
Heyo! Don't mean to be a bother. I'm really just looking for a generic female english adult voice, say for a short audiobook. Just want to use my intonation to improve the tone of text to speech really
warm stone
hi can somebody help me? I keep getting errors when I try to import my audio and then when I try to convert it, it just says error.
warm stone
like i get this kind of error
ElevenLabs_2024-03-07T06_29_51_Zayne_ivc_s50_sb75_se0_b_m2.mp3->Traceback (most recent call last):
File "D:\tin\RVC\Mangio-RVC-v23.7.0\infer-web.py", line 316, in vc_single
audio_opt = vc.pipeline(
NameError: name 'vc' is not defined
coral frigate
hi the programme works and sounds great through other applications but when i run it into discord there is alot of popping that occurs. ive gotten rid of echo and changed some other settings but still no luck. can somebody help?
dusty bramble
While training (locally) is there a way to increase the max epochs without restarting?
rare gobletBOT
Ayo? @dusty bramble level 1 !!!
eternal swift
can someone help when i do step 2 downloading the model i get this and i can continue
rare gobletBOT
Ayo? @eternal swift level 1 !!!
eternal swift
wise estuary
I trained a model over the night, every time i looked it said suc
Now i stand up and go to my pc to see after the model and it said it’s finished but i dont have any index file and 2 path files wich booth not work with wokada
Is there anything i missed?
Update: The 2 path files i mentioned above are named G_xxxxxx and D_xxxxxx the x resembles number gibberish. Both files are around 350 and 600 kb big.
Now i looked through the files and found in the weight folder ava.path, (thats my model name) and it’s something around 60 kb big.
Does anyone know where i find the index file?
proper shale
old colab, use the newer ones in https://aihubdocs.github.io/en
which also includes guides
eternal swift
thank you bro
proper shale
I trained a model over the night, every time i looked it said suc
Now i stand up...
you only need the .pth from weights
no need to upload index in realtime
proper shale
thank you bro
you're welcome! lmk if it works
eternal swift
you're welcome! lmk if it works
yo i do not know how to use this is everything i need in the how to make ai tutorials thing?
proper shale
yo i do not know how to use this is everything i need in the how to make ai tuto...
yeah, has everything ya need to know
from making ai covers to training models
eternal swift
alright thank you
proper shale
yw :)
brittle wing
i have amd rx6600 with i5-12400f which rvc version should i download?
my gpu isnt getting detected in normal version
tawny shell
will overtraining cause high buffer ms?
rare gobletBOT
Ayo? @tawny shell level 1 !!!
tawny shell
zinc anchor
It's not due to overtraining in that case, it's the chunk size you selected
proud elbow
try extra 8k-16k
tawny shell
trying to compensate with extras yeah
proud elbow
131k is too heavy
zinc anchor
Yeah I think desuwa is right
mystic ingot
What should I do?
flint vapor
What should I do?
do you have errors in the cmd?
mystic ingot
??
sand iris
??
don't put it in program files
it can't write anything in there
and the program file folder contains a space which it really hates
mystic ingot
don't put it in program files
thanks ,its works
azure marshBOT
RVC Guides (How to Make AI Cover)
wild bobcat
i have a question,
do ai voices also the ability to sing higher notes than what the input was
like if i have a sample with a deep voice, would the ai voices like pitch correct themselves to sing higher notes
low shard
error on step 2
!howtoask
patent trellis
low shard
i have a question,
do ai voices also the ability to sing higher notes than what ...
you could change the pitch to make it sound higher or deeper than the input
low shard
so the ai voice pitch corrects itself
it depends by the pitch of the ai voice model, and the input, so it can be different, but anyways you can change the pitch manually
wild bobcat
ok
another question i have is: why cant i train models anymore
i never had a nvidia gpu and trained models before just fine with normal rvc and rvc disconnected
forest current
Hey,
I can't seem to find a clear answer on this.
After training i get 2 index files, a added_xxx_flat_nprobe_ and a trained_xxxx_Flat_nprobe_ file.
Which one is the best index to use?
the added index file is always bigger, so i assume that is the correct one?
drowsy pilot
can someone please help me
ive been trying to figure out for hours
trying to make a ai voice
trim granite
why Ilaria rvc error error error
queen escarp
old colab, use the newer ones in https://aihubdocs.github.io/en
tysm ❤️😊
low shard
google colab is a service that lets you run this code with a free daily gpu (around 4 hours for thise tasks daily), its because you finished the gou and prolly connected without it, you can either wait 12-24 hours or use another acc
wild bobcat
google colab is a service that lets you run this code with a free daily gpu (aro...
oh aw man
okay
drowsy pilot
Does someone have some free time to help me out?
hot lagoon
About what?
low shard
why Ilaria rvc error error error
!howtoask
drowsy pilot
About what?
i need three ai voices for a song of like 3:20
so i just need to know how to make your own voice
i know how to make a ai cover but i just need the zip of the ai model somehow
is there a easy way or do i need to spend like weeks trying to learn it
hot lagoon
To make a ai voice?
drowsy pilot
yes
i need a voice to put over a song
a couple voices but the same song
u have a model maker role maybe you can help me out
i tried this but keep getting error in the middle right
hot lagoon
i need a voice to put over a song
I recommend using this guide to help you. It talks about overtraining, clean dataset and how to make AI voices https://aihubdocs.github.io/en
trim granite
ok i use Ilara RVC
at last step after adding model and voice file it starts processing like the screenshot below, after a while it comes up ERROR why
and sometimes it works but most of a time error error
drowsy pilot
I recommend using this guide to help you. It talks about overtraining, clean dat...
but the guide is so hard
what is inference
rare gobletBOT
Ayo? @drowsy pilot level 3 !!!
hot lagoon
what is inference
Basically, just you converting audio to the AI audio you want to have
drowsy pilot
thats what i want right
i got like very high quality audio of a dude talking
and i need his voice as singer
hot lagoon
I mean, you can also just commission a model master to make the voice for you
I see them going for like 25 euros
But that’s depend on the model, master making the voice for you
drowsy pilot
yeah but 25 is way too much
i only need the voices just for a single song of just 3:20
after that basically never
its just for a joke
hot lagoon
Well, I still recommend the guide, as it’s a very detail on what to do when making a voice model
warm flicker
16 seems pretty fine
fine deez
trim granite
has anyone tried mrbeasts deez nuts chocolate
low shard
ok i use Ilara RVC
at last step after adding model and voice file it starts pr...
are you using the google colab or hugging face space?
trim granite
hugging face
odd shale
has anyone tried mrbeasts deez nuts chocolate
What?
hot lagoon
I wonder what happened to medleyvox if anyone could provide a link to it that would be appreciated
low shard
hugging face
you need to duplicate the space then, click the top right corner , click the 3 dots, then duplicate it
odd shale
I wonder what happened to medleyvox if anyone could provide a link to it that wo...
MedleyVox isn't that good.
low shard
But i remind you that the hugging face space has no daily limitbut is way slower bc it runs on cpu
you can also use the ilaria rvc ui google colab which runs on the 4 hours free daily gpu and is faster
odd shale
MedleyVox isn't that good.
BVE V2 is better
odd shale
you can also use the ilaria rvc ui google colab which runs on the 4 hours free d...
Wasn't that colab broken?
hot lagoon
BVE V2 is better
What’s the difference between V2 and V1 of bve
odd shale
What’s the difference between V2 and V1 of bve
BVE V2 is a updated version of V1
low shard
Wasn't that colab broken?
ilaria rvc ui google colab? Nope it works, but idk why for some people sometimes it gives a “no module torch found” but i think you just need to disconnect runtime and do it again ig
hot lagoon
BVE V2 is a updated version of V1
But isn’t the two version not out yet for UVR though?
odd shale
BVE V2 is a updated version of V1
So it can isolate leads and backs better
odd shale
But isn’t the two version not out yet for UVR though?
Nope.
hot lagoon
Wait, so it is out for UVR the bve 2 cause I’m kinda little bit confused
ionic spruce
JSONDecodeError Traceback (most recent call last)
<ipython-input-12-baa61050bbe7> 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)
this error keeps up when downloading the model
trim granite
you need to duplicate the space then, click the top right corner , click the 3 d...
thanks buddy
low shard
thanks buddy
yw
sand iris
JSONDecodeError Traceback (most recent call last)
<ipy...
old colab, use a new one
-colab
sand iris
"Ilaria RVC" is most likely what you'll need
ionic spruce
"Ilaria RVC" is most likely what you'll need
thanks
trim garnet
does anyone know how I get these models ? kikiv1 index and pth
drowsy pilot
can someone please tell me what this error means and how to fix it
what am i doing wrong
it says something with like missing pth file
trim garnet
I cant uplod pic rip
odd shale
I cant uplod pic rip
@sand iris Give him image perms
low shard
also what colab are you using
sand iris
done
drowsy pilot
also what colab are you using
idk ''easy'' i think
odd shale
I cant uplod pic rip
Now you can
trim garnet
Now you can
thankiu
trim garnet
You're welcome.
I saw this on youtube but rip he didnt said anything about these models
drowsy pilot
does anyone know a website where u can cut a video into shorts?
so like 10 minute video in pieces of 9 seconds
glad zealot
so like 10 minute video in pieces of 9 seconds
i think you can do that with ffmpeg locally but idk a website to do that
low shard
idk ''easy'' i think
could you send the link just to check up?
drowsy pilot
odd shale
does anyone know a website where u can cut a video into shorts?
Uhhm, you can use Audacity, Free Batch Music Splitter or FFMPEG
low shard
i tried this one
https://colab.research.google.com/drive/1r4IRL0UA7JEoZ0ZK8PKfM...
okay good its the updated one, dw just wanted to check as some confuse it with the outdated one
drowsy pilot
ugh im getting so fricking frustrated
i feel really bad
i just want to get this over with
why is it so damn complicated to make a voice
isnt there a site to just yeet a audio file in and it will just replace the song
low shard
isnt there a site to just yeet a audio file in and it will just replace the song
nope
you can either make a #1159289738314919936 , or make it yourself
it takes a bit of patience
drowsy pilot
if i request a model nobody will ever make it
i just want to give my bro the song that he asked for
i got the audio files and everything
just need the voice
low shard
if i request a model nobody will ever make it
i mean, you could either wait or request a paid one
or have patience and make it, it gets easier with practice and time
honest jetty
how do i deal with this error
rare gobletBOT
Ayo? @honest jetty level 1 !!!
honest jetty
python3: can't open file '/content/src/main.py': [Errno 2] No such file or directory
2
help pls
ionic spruce
is it normal for the conversion to take more than 3 minutes?
rare gobletBOT
Ayo? @ionic spruce level 1 !!!
low shard
is it normal for the conversion to take more than 3 minutes?
i mean, depends by the audio input lenght and if its on cou, what are you using?
low shard
python3: can't open file '/content/src/main.py': [Errno 2] No such file or direc...
what tool u using?
honest jetty
CoverGen_NO_UI_v2_en.ipynb
ionic spruce
i mean, depends by the audio input lenght and if its on cou, what are you using?
Applio, by IA Hispano Huggingface Spaces
low shard
Applio, by IA Hispano Huggingface Spaces
yea hugging face spaces run on cpu and has no daily limits (be sure you clicked the 3 dots at the top right corner, duplicate the space) so its wayy slower, you could just wait or use the applio ui colab which is 4 free hours daily of gpu and its way faster
ionic spruce
yea hugging face spaces run on cpu and has no daily limits (be sure you clicked ...
ok thanks a lot
hard gust
Somebody tried this on AMD? https://github.com/vosen/ZLUDA
low shard
heres the tool
In which step did you get that error?
low shard
ok thanks a lot
Yw
honest jetty
In which step did you get that error?
generating the cover
as soon i click the play button in the blue circle, that error occurs
low shard
as soon i click the play button in the blue circle, that error occurs
Sorry but i dont know much of ai cover gen colab, maybe @viscid moss could help
drowsy pilot
i really dont understand how people like websites like github or huggingface
worst websites ever
sand iris
i really dont understand how people like websites like github or huggingface
what are you having issues with ?
drowsy pilot
uploading ai model
sand iris
oh
drowsy pilot
huggingface is giving me flashbacks from github when i did an IT study
both huggingface and github are terrible
really annoying websites
why does it all have to be so complicated
how do i upload something to use a voice in huggingface?
can someone please help me
hello?
rare gobletBOT
Ayo? @drowsy pilot level 4 !!!
granite sapphire
So if I am making/recording a model of myself. I've seen "about 10 mins is fine" obviously more can allow for more 'dataset' but curious if normal spoken word (Reading scripts etc.) Is better for registering vocals. Or if singing/more differentiating vocals is better for a dataset?
drowsy pilot
cant someone please just help me
why is this chat so dead man
i just need a little bit of help
ive been trying for like 4 hours
granite sapphire
ive been trying for like 4 hours
What are you trying to do exactly? Are you just trying to do simple AI covers w/ already done voice models?
rare gobletBOT
Ayo? @granite sapphire level 1 !!!
drowsy pilot
What are you trying to do exactly? Are you just trying to do simple AI covers w/...
no making a ai cover with a newly made voice
but i need a huggingface link to put in any ai cover machine thng
granite sapphire
but i need a huggingface link to put in any ai cover machine thng
Where did you make the model?
Did you do it off of RVC Mainline/offline or?
drowsy pilot
in some random google collabs thing
granite sapphire
Ah for sure, I haven't messed with a colab training thing, but did it give you any physical files to download?
drowsy pilot
yes
like 10 minutes of mp3
i have a pth file now
it should be the model
i just want to see if it works
granite sapphire
i have a pth file now
Do you have 7zip?
Did it also give you an index file? (index is optional technically for most programs)
But requires for huggingface AFAIK
drowsy pilot
to be fair i have no idea what the heck all of that means
i just followed an tutorial and got some files now
granite sapphire
i just followed an tutorial and got some files now
For sure I'm just trying to gauge what you all have.
Basically in order for you to make an "AI cover" I reccommend for a striaght forward experience AI Covergen (You can find the link in the server somewhere it pops up in General from time to time) go to that Colab. run all the modules until it brings you to the interface.
As far as "preparing" your AI model. Download 7zip (Its free, google it.) Then gather at least the .py file and possibly the index file, highlight them both and right click, mouse over 7zip and click "compress as zip file" then it will create a zip file for your model.
Then when you navigate back over to AI Covergen, go to the upload model tab, drag the zip file into the upload model box, give it a name and click "upload" wait for it to say "Model uploaded!"
Go back to the main page, click refresh index/models. Click your model from the drop down. Then with AI Covergen all you have to do is copy a youtube link and paste it in, then click run. It'll do the rest for you if you just want to do some quick tests
brittle wing
i have amd rx6600 with i5-12400f which rvc version should i download?
my gpu isnt getting detected in normal version
viscid moss
python3: can't open file '/content/src/main.py': [Errno 2] No such file or direc...
It can be 2 things:
- You didn't run the cells in order
- Or maybe the error appeared randomly
In any case you can restart the execution environment and start again
viscid moss
i have amd rx6600 with i5-12400f which rvc version should i download?
my gpu isn...
For voice changer?
brittle wing
For voice changer?
ye ye sir
rare gobletBOT
Ayo? @brittle wing level 1 !!!
brittle wing
i want to make a song in my own voice model
i cant make a voice model as my gpy isnt getting detected
sand iris
i cant make a voice model as my gpy isnt getting detected
can't train on amd gpus
sand iris
relies on google servers
You'll be able to (ab)use one of their VMs with a GPU for a few hours, enough to train a model
upper furnace
Help
ocean socket
tryna use "Mangio-RVC-v23.7.0" for ai cover, I can't use the easy gui (getting path not found error) and standard webpage has a weird auto-translation to my own language which somehow makes it even harder to understand the gui, is there a known solution for the easy gui or a way to disable translation?
foggy folio
guys is this ok for uvr
rapid beacon
ig try removing aria2 from the code then do !pip install aria2 after
Just like how you wrote it? 😧
summer mirage
always when i try i Get this error Traceback (most recent call last):
File "D:\Downloads\aivoicetrainer\runtime\lib\site-packages\gradio\routes.py", line 321, in run_predict
output = await app.blocks.process_api(
File "D:\Downloads\aivoicetrainer\runtime\lib\site-packages\gradio\blocks.py", line 1007, in process_api
data = self.postprocess_data(fn_index, result["prediction"], state)
File "D:\Downloads\aivoicetrainer\runtime\lib\site-packages\gradio\blocks.py", line 953, in postprocess_data
prediction_value = block.postprocess(prediction_value)
File "D:\Downloads\aivoicetrainer\runtime\lib\site-packages\gradio\components.py", line 2076, in postprocess
processing_utils.audio_to_file(sample_rate, data, file.name)
File "D:\Downloads\aivoicetrainer\runtime\lib\site-packages\gradio\processing_utils.py", line 206, in audio_to_file
data = convert_to_16_bit_wav(data)
File "D:\Downloads\aivoicetrainer\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'
Traceback (most recent call last):
File "D:\Downloads\aivoicetrainer\my_utils.py", line 14, in load_audio
ffmpeg.input(file, threads=0)
File "D:\Downloads\aivoicetrainer\runtime\lib\site-packages\ffmpeg_run.py", line 325, in run
raise Error('ffmpeg', out, err)
ffmpeg._run.Error: ffmpeg error (see stderr output for detail)
During handling of the above exception, another exception occurred:
half cove
Just like how you wrote it? 😧
hm?
summer mirage
always when i try i Get this error Traceback (most recent call last):
File "D:...
@half cove do you know that error
half cove
<@936673139419664414> do you know that error
show me ur file path
summer mirage
what do you mean?
summer mirage
mangio?
half cove
wait u aint the same guy as before
summer mirage
with all the voices i get the same error
minor zodiac
Help, I'm starting to lose my mind "FileNotFoundError: [Errno 2] No such file or directory: 'C:\Mangio-RVC\Mangio-RVC-v23.7.0/logs/Spanish/preprocess.log'"
And before you ask, no. It does not contain any spaces
@tawdry ruin
Help?
cold raptor
Help, I'm starting to lose my mind "FileNotFoundError: [Errno 2] No such file or...
I think you don't have file
cold raptor
minor zodiac
The model name should also not have spaces
cold raptor
Np
minor zodiac
cold raptor
I helped you a lot
cold raptor
Np
bleak smelt
I keep getting an error when trying to download my model, it says it cannot access the gdrive link even though it is set to public
rare gobletBOT
Ayo? @bleak smelt level 2 !!!
stone pagoda
hello. I installed rvc webui but I couldn’t select inference model. do you know why I couldn’t select inference model on webui?
azure marshBOT
RVC Guides (How to Make AI Cover)
drowsy pilot
What type of audio does work best for making a model?
as in talking or singing
whispering, shouting etc
also is more epochs better?
proper shale
What type of audio does work best for making a model?
depends on what you wanna achieve
proper shale
whispering, shouting etc
shouting is just generally not recommended cuz it messes up the normal voice
drowsy pilot
the person isnt a rapper so talking it is then i guess
proper shale
understandable
coral frigate
hi the programme works and sounds great through other applications but when i run it into discord there is alot of popping that occurs. ive gotten rid of echo and changed some other settings but still no luck. can somebody help?
dusty bramble
How many epochs for 40 minutes of data?
ocean socket
how do i upload my own voice model to applio
mental lantern
can you use the rvc to play an mp3 file?
brittle wing
whats the best website for ai voices i need dogday's voice please help
proper shale
hi the programme works and sounds great through other applications but when i ru...
probably the virtual audio cable's fault - if you're using the one by VB Cable then try VAC by Muzychenko
proper shale
How many epochs for 40 minutes of data?
go for 500, every dataset length
proper shale
can you use the rvc to play an mp3 file?
if you're using W-Okada just set the input to a file. but honestly id just recommend installing a local fork for that
dusty bramble
go for 500, every dataset length
Shit, I went for 250
Next one I try I'll do 500
proper shale
Shit, I went for 250
Next one I try I'll do 500
Oh that's fine
dusty bramble
I barely know how to make it work, if I can keep training it up to 500 I'll do that
proper shale
If it doesn't overtrain you can continue to 500, then 750, then 1000 and so on
dusty bramble
Yeah, that's I'm scared of
That's why I was asking about dataset length and epoch
I did 250 just cuz I'm training on a laptop and wanted to get it done with it lol
proper shale
most datasets don't really overtrain at 250, especially on higher batch sizes
but ofc, every model's different
brittle wing
can't train on amd gpus
how much time it takes to train it
can i pay you if you can train my voice?
sand iris
how much time it takes to train it
I don't really know, haven't used colab myself
sand iris
can i pay you if you can train my voice?
I don't do that But ig you can go to #1159289738314919936
drowsy pilot
is it better to insert short audio segments to make a ai model or does it not matter at all
ornate badger
So, what am I supposed to do to connect with it? I hope Google doesn't do anything shady to my stuffs.
I want to use the RVC, but I feel unsure about Google forcing ths to me.
I need help
shy karma
guys some ai covers vocals can not catch up the original vocal rhytm. any solutions?
ornate badger
So, what am I supposed to do to connect with it? I hope Google doesn't do anythi...
Which I should connect with?
reef violet
-rvc
candid crescent
What does S.R do and should I have it on high?
earnest grove
guide
stone skiff
hey guys, I'm using the tutorial in this video, and found this comunity. How can I generate audio with people saying things in Portuguese Brasil? https://www.youtube.com/watch?v=7tpWH8_S8es
I have some experience in Stable diffusion but I don't see any place where I can change the "model"? I'm assuming I should have a model in portuguese brasil and use it?
low shard
So, what am I supposed to do to connect with it? I hope Google doesn't do anythi...
Yea connect it, google colab is a service of google and thats where the colabs of rvc, it just connects to google drive the time while the code is running to auto save the model files if you are making a model for examples
ornate badger
Yea connect it, google colab is a service of google and thats where the colabs o...
I have very hard time using Aplio. How do you even load the voice model into it
low shard
What colab are u using btw
rare gobletBOT
Ayo? @ornate badger level 2 !!!
ornate badger
low shard
What does S.R do and should I have it on high?
Thats wokada ai voice changed, please ask in #🔍│help-w-okada or #1192011222023950368
candid crescent
Thats wokada ai voice changed, please ask in <#1159290161683767298> or <#1192011...
🫡
ornate badger
Check this https://aihubdocs.github.io/en/rvc/cloud/applio-colab/#inference-
I did the 1st step.
But the 2nd step doesn't work. No models shown.
half cove
I did the 1st step.
But the 2nd step doesn't work. No models shown.
after refreshing model does it still not show up?
low shard
I did the 1st step.
But the 2nd step doesn't work. No models shown.
Click refresh button next to it
ornate badger
after refreshing model does it still not show up?
I did refresh, and still nothing
half cove
if it doesnt then show what error colab run gui cell says
ornate badger
Click refresh button next to it
It doesn't work
ornate badger
if it doesnt then show what error colab run gui cell says
Where do we see it?
low shard
Where do we see it?
Go in the applio ui colab page, check the output of the cell to run ui
ornate badger
Go in the applio ui colab page, check the output of the cell to run ui
Which?
low shard
Which?
no i mean like, in the colab page, the one where you runned the install and run cells
ornate badger
no i mean like, in the colab page, the one where you runned the install and run ...
Not sure where it is
low shard
Not sure where it is
before having that applio interface, what did you do?
ornate badger
before having that applio interface, what did you do?
Well running the Applio Rvc
ornate badger
Then go to the link
low shard
Well running the Applio Rvc
yea, check the page that runned it
ornate badger
1st step, I go to Apllio collab
Then I click on run all
It has the tensorboard thing which I have no idea how this work
I wonder if there is another model easier to use than this one.
dawn nebula
It has the tensorboard thing which I have no idea how this work
an indicator if your models overtrained or not
ornate badger
an indicator if your models overtrained or not
I guess my links from Google drive couldn't even be loaded.
I tried one from Huggingface and it worked
mint linden
So, what am I supposed to do to connect with it? I hope Google doesn't do anythi...
it will hack your computer and explode your gpu
(trust me i developed rvc)
ornate badger
it will hack your computer and explode your gpu
How do you disconnect?
rare gobletBOT
Ayo? @ornate badger level 3 !!!
mint linden
How do you disconnect?
you don't.
now you have like 20 viruses
you're screwed man.............
actually no you have nearly 600,000 trojans aswell as a nuclear bomb microchip in your hard drive
ornate badger
You should have told sooner
mint linden
You should have told sooner
jokes aside it only checks for the rvcdisconnected folder
it won't cause anything nor hack your shit since it can't look through your files
half cove
it won't cause anything nor hack your shit since it can't *look* through your fi...
cant joke with ai hub members they will believe it 😭
for a long dataset you might wanna use a higher batch size but its better just to shorten the dataset by only keeping the best bits
rare gobletBOT
Ayo? @viral comet level 2 !!!
dawn nebula
I guess my links from Google drive couldn't even be loaded.
did u pack ur stuff properly
ornate badger
did u pack ur stuff properly
Not sure what you meant
dawn nebula
so u wanna train or infer
ornate badger
so u wanna train or infer
I just wanna make a voice sing a song or say something
dawn nebula
I just wanna make a voice sing a song or say something
show me what u put inside ur google drive link
ornate badger
show me what u put inside ur google drive link
zips of voices
I had no choice but do the manual way, It works
low shard
cant joke with ai hub members they will believe it 😭
Fr😭
dawn nebula
you can only import 1 at a time
tulip sierra
After I deleted the AI, my audio started to fail on my PC and my games and YouTube videos started to freeze, do you know what it could have done to cause these problems on my PC? What should I do?
shrewd apex
Hi, I'm getting a: No module named 'utils' and 'RVC' object has no attribute 'tgt_sr' when trying to run a RVC I found here
Could someone point me in the right direction, please?
dawn nebula
Hi, I'm getting a: No module named 'utils' and 'RVC' object has no attribute 'tg...
which rvc are u trying to run
odd shale
Hi, I'm getting a: No module named 'utils' and 'RVC' object has no attribute 'tg...
Which RVC version are you trying to use?
Locally or Colab?
shrewd apex
I don't know what that means, sorry. What I do know is that it works fine with the bundled .pth files but I got that error when trying to run one from here. https://discord.com/channels/1159260121998827560/1214406843590377482 Here is the one I'm trying to run.
low shard
I don't know what that means, sorry. What I do know is that it works fine with t...
Thats an gpt so vits model
it doesnt work for rvc
shrewd apex
Oh ok. Thank you, how can I make sure that I download rvc models?
low shard
Oh ok. Thank you, how can I make sure that I download rvc models?
Be sure to always look at the tags of the voice models in #1175430844685484042
there is a tag for rvc models
and one for gpt so vits ones
you can try another rvc model
shrewd apex
What software do I need to run vits?
rare gobletBOT
Ayo? @shrewd apex level 1 !!!
gusty kestrelBOT
@kindred jungle, I have found 1 results that match your search!
Requested by s_sayurii.
rustic shale
I had a question, is 25 mins of speech good for a dataset for traning RVC?
Or is that too much
glad zealot
I had a question, is 25 mins of speech good for a dataset for traning RVC?
That should be good
glad zealot
What software do I need to run vits?
You need to use GPT-SoVITS, yes that's the name of the program
shrewd apex
You need to use GPT-SoVITS, yes that's the name of the program
Thank you
rustic shale
That should be good
thank you
rare gobletBOT
Ayo? @rustic shale level 1 !!!
azure marshBOT
RVC Guides (How to Make AI Cover)
dusty bramble
If it doesn't overtrain you can continue to 500, then 750, then 1000 and so on
using mangio rvc training, how do i pick up on training a voice?
like, i finished training but i want more
proper shale
like, i finished training but i want more
just insert the same settings as before, no need to preprocess or feature extract again, use same batch size, and it should do it
you can change save frequency and total training epochs though
proper shale
but i need to redo from scratch again?
nope
just the same settings
and it should work lol
dusty bramble
so i put the file name there and start?
proper shale
so i put the file name there and start?
nah keep that as is
brittle wing
Hi, anyone has a collab link?
brittle wing
Thanks
proper shale
you're welcome!
dusty bramble
oh, yeh got it, ty
rare gobletBOT
Ayo? @dusty bramble level 2 !!!
proper shale
oh, yeh got it, ty
lmk if it works!
dusty bramble
worked, ttyty
old oracle
anyone got a hello kitty model 😭
spare shale
anyone got a hello kitty model 😭
But I think these are bad models
You can find any model here: https://www.weights.gg/
trim sparrow
anyone has a google colab link works with gradio ? I have a paid colab account
spare shale
But I think these are bad models
You can find any model here: https://www.weigh...
@hollow gulch !
You can find any model here: https://www.weights.gg/
molten pecan
anyone has a google colab link works with gradio ? I have a paid colab account
Ilaria RVC, Applio Colab, AICoverGen UI or EasyGUI
-colab
turbid meteor
what is the pitch detection alg?
viscid moss
what is the pitch detection alg?
azure marshBOT
RVC Guides (How to Make AI Cover)
dusty bramble
I'm sorry for coming here frequently but what does "max value tensor (random number here)" mean
sand iris
I'm sorry for coming here frequently but what does "max value tensor (random num...
you're using the real time voice changer ?
Nothing more than debug stuff that managed to make it's way there for some reason
dusty bramble
No, trying to train a model
dusty bramble
Nothing more than debug stuff that managed to make it's way there for some reaso...
Correcting myself a bit
It's actually "max value is tensor (random number)"
sand iris
Same thing applies as far as I remember
dusty bramble
So, just ignore it?
sand iris
Yeah
dusty bramble
Okok thank u
strange burrow
What's a good software for voice modeling?
rare gobletBOT
Ayo? @strange burrow level 1 !!!
brittle wing
Okay hear me out
I trained the model and switched about 2-3 accounts so I can finish the task earlier now I created a new account and I tried to reach 500 epochs but when I try to resume it starts from 1 and it pisses me off
rare gobletBOT
Ayo? @brittle wing level 3 !!!
brittle wing
I don't understand
sand iris
I'm not so sure about that anymore
what about it ?
dusty bramble
what about it ?
Only trains one epoch then keeps just spamming that
dusty bramble
Rtx 3060 laptop
Batch size is 3
I was able to train a voice locally up to 350 epoch before
rare gobletBOT
Ayo? @dusty bramble level 3 !!!
ocean socket
-colab
dusty bramble
@sand iris I think it's a dataset issue
Which I don't understand why it isn't working
Got the data set the same
sand iris
I don't do much model training, but it shouldn't be able to silently fail like that
except if it's giving out anything related to NaN
dusty bramble
I'm gonna try to deal with it tomorrow
Thank u still tho
brittle wing
I don't do much model training, but it shouldn't be able to silently fail like t...
Why did it fail with me
sand iris
I really don't know
brittle wing
-rvc
brittle wing
-help
brittle wing
-overtrain
azure marshBOT
All-In-One Guide on how to make a good model
This guide explains how the D and G files works and much more: https://rentry.org/RVC_making-models
Credits: LUSBERT
Automated Overtraining Detection (AOD)
Will be available soon in #1159513888199540817
Credits: grvyscale
brave garnetBOT
Made with love by Ilaria 💖
⠀
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 💖
viral sentinel
hey im new to rvc model making and its my first attempt but my tensorboard shows up with this error what should i do
low shard
hey im new to rvc model making and its my first attempt but my tensorboard shows...
you need to click on active and choose scalars instead and refresh, check https://aihubdocs.github.io/en/rvc/resources/epochs-overtraining--tensorboard/
upbeat coral
how long does 800 epoch take on average?
mine said 7 hours eta
wise estuary
how long does 800 epoch take on average?
mine said 7 hours eta
I have a RTX 4080 and did like 200 epochs per hour
upbeat coral
I have a RTX 4080 and did like 200 epochs per hour
okay. I only have a 3070. am looking at possibly getting a 4070
upbeat coral
If you close out of web browser but leave the RVC up in the terminal up, does it still go?
wise estuary
If you close out of web browser but leave the RVC up in the terminal up, does it...
i just didnt touch anything while i did the model
latent kettle
i just didnt touch anything while i did the model
hey can you train a deepfake model for me ?
wise estuary
hey can you train a deepfake model for me ?
Uhhhh what is that? So like those video things?
latent kettle
ahh not really
wise estuary
ahh not really
You mean a voicemodel?
I can do that if you want, but i dont want to search for the audio stuff
latent kettle
You mean a voicemodel?
let's talk in DM ?
wise estuary
Ok
eternal swift
how hard is training a voice model. im looking at the tutorial and it seems like on my shitty laptop that it would take forever. but the current ai sounds kinda robotic
latent kettle
how hard is training a voice model. im looking at the tutorial and it seems like...
it's easy
proper shale
how hard is training a voice model. im looking at the tutorial and it seems like...
if you're gonna use colab it won't take THAT long
of course you have to limit the training session but like... depending on the dataset you can get 100 epochs after an hour, which already should sound decent
undone valve
hi guys, what's the best solution to train a voice model with an AMD GPU?
I tried RVC V2 on collab yesterday but it seems that i'm now restricted 😔
proper shale
hi guys, what's the best solution to train a voice model with an AMD GPU?
I tri...
just try the recent colabs, like rvc disconnected:
https://colab.research.google.com/drive/1XIPCP9ken63S7M6b5ui1b36Cs17sP-NS
if you want guides: https://aihubdocs.github.io/en
undone valve
just try the recent colabs, like rvc disconnected:
https://colab.research.googl...
thanks, that's the one I'm using and the one that got me restricted 😅 I hope it's only temporary
molten pecan
just try the recent colabs, like rvc disconnected:
https://colab.research.googl...
hey MJ
fading drum
How do i clone a voice using rvc? (https://huggingface.co/lj1995/VoiceConversionWebUI/blob/main/RVC1006Nvidia.7z) im not getting it
rare gobletBOT
Ayo? @fading drum level 1 !!!
fading drum
because i dont have a .pth but i have a dataset with about 90 audio files about 10 secs in lenght, do i NEED a pth file to initiate it?
molten pecan
because i dont have a .pth but i have a dataset with about 90 audio files about ...
- in order to create a model from scratch, you don't need the .PTH file, just the dataset & RVC
- the dataset doesn't need to be split nowadays. 1 audio file is enough
https://aihubdocs.github.io/en/rvc/local/mainline/ here's a guide for using Mainline RVC =3
sand kelp
Hi. I'm a complete beginner, trying to complete an assignment to make an AI cover. I'm trying to get the Google Collab to work. Step 1 is fine, but step 2 always gives me an error. I'm using this AI model link: https://drive.google.com/file/d/1FBZhVB9Py0cUmP_jsgvve8QiL-B3DCnL/view?usp=sharing. The result I get is below. Any help is appreciated!!!
JSONDecodeError Traceback (most recent call last)
<ipython-input-5-a9f869d1cf93> 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)
proper shale
thanks, that's the one I'm using and the one that got me restricted 😅 I hope it...
oooh that's it then. yeah it's temporary
proper shale
Hi. I'm a complete beginner, trying to complete an assignment to make an AI cove...
old colab, use more recent ones
https://aihubdocs.github.io/en has guides and newer colavs
proper shale
...looks ok
just add the 2333333 thing
cause that's what happens when you only save latest ckpt
molten pecan
just add the 2333333 thing
nah I don't think that's necessary. that's a RVC-D thing only
I mean like, out of all the software explained in the docs, RVC-D has it activated by default
gusty kestrelBOT
@fallow crescent, I have found 25 results that match your search!
@fallow crescent, I have found 25 results that match your search!
Requested by hades1989.
low shard
@fallow crescent theres #🔍│find-models to use the /search commands, just to not spam up the help channels yk
molten pecan
<@635908828734554122> theres <#1163592055830880266> to use the /search commands,...
morning Nick
low shard
morning Nick
goodmorning julia
low shard
so the voice.ai app is locked behind a paywall right?
voice.ai is a paid software not affilied with us
we use wokada ai voice changer here
which is all free
ask in #🔍│help-w-okada
low shard
does it consume a lot of gpu?
well depends on your pc, and if your pc isnt good enough you can use the online colab version
drifting tendon
m
rare gobletBOT
Ayo? @drifting tendon level 1 !!!
drifting tendon
k
low shard
the right channels to talk abt it are #1192011222023950368 or #🔍│help-w-okada as this is an help channel for rvc, which isnt a realtime ai voice changer
drifting tendon
thanks
low shard
yw
proper shale
nah I don't think that's necessary. that's a RVC-D thing only
no it isn't lol
candid crescent
is the RVC help missing a step or am I looking at the wrong thing? https://aihubdocs.github.io/en/rvc/local/mainline/
Its telling me to get a path file, but so far it's only told me to get sound, and clean the audio.
(Last tip)
molten pecan
is the RVC help missing a step or am I looking at the wrong thing? https://aihub...
What are you trying to do exactly?
Train a model?
candid crescent
What are you trying to do exactly?
Yes.^
rare gobletBOT
Ayo? @candid crescent level 4 !!!
candid crescent
This.. is the how to make voice models.
The last tip it gave me is as you see above
Then it's just completely skipped to "
"
molten pecan
Then it's just completely skipped to "
Uhh I'm confused. I checked the three pages and they have the accurate information
candid crescent
Am I missing some?
molten pecan
https://aihubdocs.github.io/en/essentials/how-to-make-voice-models/ this is the starting point
candid crescent
Right and you click on how to prepare dataset
And sound isolation is the last thing available
molten pecan
And sound isolation is the last thing available
Uhh nope, after cleaning the dataset the page tells you to set up RVC
ashen sleet
Hi friends, I am just trying to run RVC for the first time, and encountering a confusing issue.
I ran install_easyGUI.bat and got the following error after lots of successful package installations:
tensorflow 2.6.0 requires numpy~=1.19.2, but you have numpy 1.23.5 which is incompatible.
tensorflow 2.6.0 requires typing-extensions~=3.7.4, but you have typing-extensions 4.10.0 which is incompatible.
onnx 1.14.0 requires protobuf>=3.20.2, but you have protobuf 3.17.3 which is incompatible.
skl2onnx 1.14.1 requires scikit-learn<1.3,>=0.19, but you have scikit-learn 1.3.0 which is incompatible.```
So I have installed versions that are too new? This is possibly from my installation of stable diffusion, at a guess. Anyway, does anyone have a suggestion for the best way to go about addressing this?In the step 2
candid crescent
Uhh nope, after cleaning the dataset the page tells you to set up RVC
Righhttt i've installed RVC
and it's immediately skipped to having .pths
am I confused or am I supposed to have a .pth for my model already
molten pecan
am I confused or am I supposed to have a .pth for my model already
Nope. The .pth is the model itself. To make a model you only need the dataset and RVC
candid crescent
Nope. The .pth is the model itself. To make a model you only need the dataset an...
ah so the .pth isn't required?
should I just continue scrolling?
molten pecan
ah so the .pth isn't required?
Yeah. RVC will make it from scratch
You don't need it beforehand
candid crescent
Yeah. RVC will make it from scratch
thanks man sorry I was really confused
molten pecan
Np. Good luck
candid crescent
this section can be ignored?
molten pecan
Hi friends, I am just trying to run RVC for the first time, and encountering a c...
Sheesh @half cove @low shard any ideas?
molten pecan
this section can be ignored?
Inference is applying a voice model to an input audio. Like making AI covers
Like the tutorial said, simply download RVC. Read the Installation chapter
https://aihubdocs.github.io/en/rvc/cloud/ilaria-rvc/#troubleshooting- check the troubleshooting
That error is pretty common dw
bleak smelt
I keep getting an error saying it cannot access my model (google drive) because its set to private but its not?
molten pecan
I keep getting an error saying it cannot access my model (google drive) because ...
How's the General Access set as?
bleak smelt
anyone with link
low shard
Sheesh <@936673139419664414> <@911742715019001897> any ideas?
sorry but nope unfortunately, idk much of local
bleak smelt
ilaria
molten pecan
ilaria
That's odd
https://aihubdocs.github.io/en/rvc/cloud/ilaria-rvc/#troubleshooting- check the troubleshooting
bleak smelt
Yeah it seemed to have been working fine a few days ago but out of nowhere it just keeps telling me that
half cove
Hi friends, I am just trying to run RVC for the first time, and encountering a c...
jus use normal
i would help but my mangio just doesnt exist
😭
ashen sleet
jus use normal
I tried that too, before, I swear, and it didn't work, I just ran go-web.bat again to get the error screenshot for you... and it worked 😂
So, thank you for somehow solving my problem with literal magic
half cove
low shard
lmao
keen silo
im using ilaria rvc to do a cover but when i downloaded a model
it did not show up
rare gobletBOT
Ayo? @keen silo level 2 !!!
bleak smelt
Cannot retrieve the public link of the file. You may need to change
the permission to 'Anyone with the link', or have had many accesses.
You may still be able to access the file from the browser:
this is the error i get
low shard
Cannot retrieve the public link of the file. You may need to change
the perm...
send the link of the model
candid crescent
I cannot get this to work. I got the exact file and it's not giving me anything to download.. am I missing something again?
bleak smelt
even tried using a different gdrive
half cove
I cannot get this to work. I got the exact file and it's not giving me anything ...
specify the file path to the output folder
candid crescent
specify the file path to the output folder
I cannot. It just refuses saying it's not a real file path
I'll show you now
oh and my whole file explorer is consistently freezing
and then this happens
half cove
just use uvr
candid crescent
just use uvr <:nails:1159569314848972891>
UVR?
half cove
also what gpu go you have
azure marshBOT
Suggestion for @candid crescent
half cove
the one from mainline rvc is poopy
candid crescent
candid crescent
the one from mainline rvc is poopy
I had already used UVR to split my audios, can I just use that straight up or does it have to be processed from this RVC
half cove
just split using uvr then use those audios
candid crescent
😭
low shard
https://drive.google.com/file/d/1BqogpWTox1So3S3xzltn5frHTFyBuvn6/view?usp=shari...
it seems to work, which ilaria rvc are you using?
bleak smelt
candid crescent
just split using uvr then use those audios
I still need RVC to train the models right?
frank isle
yo quick question, what's the recommended duration of the samples for the dataset?
low shard
this one: https://colab.research.google.com/drive/16LkwvFZeudTpUOsE_6bMjOq2qkxFo...
oh yea the colab, i just tried that, it works fine
be sure the name you gave it has no spaces
bleak smelt
@low shard yeah the name has no spaces, idk whats up with it, because it'll say its downloaded, but never pops up
low shard
<@911742715019001897> yeah the name has no spaces, idk whats up with it, because...
click runtime, disconnect runtime and delete, then re run everything again and retry to download it
bleak smelt
click runtime, disconnect runtime and delete, then re run everything again and r...
tried that too
low shard
tried that too
could you send a ss when it shows the error?
wise estuary
Can i pause the model creation?
low shard
tried that too
wait nvm yea i just seen, the colab gives this error to me too
bleak smelt
yeah idk whats up with it
candid crescent
but both options are 40k to 48k
No answers?