RIP ELEVENLABS! Make THE BEST Text-To-Speech Locally!
Added 2024-05-09 16:58:14 +0000 UTCHey everyone! I've just released a new video on how to make the best Ai text-to-speech voices on your local computer with 3 different methods and 3 Webui! So that no matter who you are and what your goals are, you get the best quality possible! Bye-bye elevenlabs!
You can check out the video right here: https://youtu.be/ds5LLIt5OLM
I also created a 1-click installer for the 3 Webui for my Patreon supporters, that you can find right here: https://www.patreon.com/posts/103930347
You can also find the free PDF guide for the methods down below 👇
As always, supporting me on Patreon allows me to keep creating helpful resources like this for the community. Thank you for your support - now go have some fun😉!
Comments
send me a dm
Aitrepreneur
2024-11-28 16:09:23 +0000 UTCHello... Unfortunately I have a problem with the installation. I always get the message at the end: xtts-webui-main\xtts-webui-main\scripts\modeldownloader.py”, line 4, in import requests ModuleNotFoundError: No module named 'requests' Press any key . . . how can i fix the error? Can anyone help me further ? Best regards: Lissy G
Lissy G
2024-11-27 02:56:24 +0000 UTCI installed specific version of numpy: pip install numpy==1.22.0 and then requirements with pip install -r requirements.txt --no-deps .. it seems to go through now ok, but I need to test it later today.
Janne Kallio
2024-11-08 17:22:29 +0000 UTCI downloaded the vs_BuildTools.exe manually and ran it. There I noticed difference to the installation with your .bat-file. when running the .bat-file, it starts the Vs BuildTools installation and asks for permission to run, but after that it seems to skip the installation. When I ran the exe-file manually, it actually install the VS tools. So this might be a real issue in the installer bat. I ran the installer for the XTTS-RVC-UI and it ran fast, but returned to the error about numpy. I installed then the numpy and ran the install requirements.txt. now it is again installing slowly and pip is trying to find correct versions. I fear it will again take hours and end with the error: pip._vendor.resolvelib.resolvers.ResolutionTooDeep: 2000000 .... can the problem be related to the version of numpy?
Janne Kallio
2024-11-08 11:17:53 +0000 UTCI still got error: " ModuleNotFoundError: No module named 'numpy'" and after "pip install numpy" and installing the "requirements.txt" I am again looking for hours at "INFO: this is taking longer than usual..." EDIT: I left it run and got ERROR in the end ... pip._vendor.resolvelib.resolvers.ResolutionTooDeep: 2000000 .. so something in my setup is causing conflicts.. investigation continues..
Janne Kallio
2024-11-07 14:36:44 +0000 UTCThanks for this and the content in general.. I am learning a lot! so much, my head hurts :) You know what would be a good idea for videos. Management, installation and maintenance of software and files. I feel like I need to search, download, install, remove, search, download, reinstall... over and over again... and I have no system for it. I will try from now on have some folder just for these AI related files and scripts, so I can easily jump between the different versions .. as it seem I will next try to setup the cuda 11.8 .. so now I add to my AI related scripts: nvcc --version .. and to the files folder both cuda_11.8.0_522.06 and cuda_12.6.2_560.94. maybe it could be possible to have a bat file that checks the current versions installed and depending on the situation uninstalls and installs the correct files from such folder?
Janne Kallio
2024-11-07 11:50:13 +0000 UTCindeed make sure that you have first, the python 3.10.11 (or 3.10.6 but now I recommend the 3.10.11 instead) and that it's added to path with no other python install, make also sure you got cuda 11.8 installed and added to path as well before launching the installer
Aitrepreneur
2024-11-07 00:31:41 +0000 UTCwhen installing the xtts-finetune-webui I start to get notes: INFO: pip is looking at multiple versions of ... to determine which version is compatible with other requirements. This could take a while. ... INFO: This is taking longer than usual. You might need to provide the dependency resolver... I am not able to get to the end of the install as it gets slower and slower.. I tried editing the installers to update pip version etc, but it just got worse ... now I removed the installed folder and trying a clean install from beginning.. and gettign the same "notes".... any idea why I am getting this issue? also how does the CUDA version affect these installs? I seem to have cuda_12.6.r12.6 and the install is using cu118.. can this be the issue? I did remove all python installs and reinstalled the recommended python310 version and added it to the path. EDIT: I managed to do the install directly from https://github.com/daswer123/xtts-finetune-webui .. but I guess I still need to fix the XTTS-RVC-UI
Janne Kallio
2024-11-06 11:11:18 +0000 UTCMake sure that the cuda12 was installed correctly and added to path in the environment variables, if it is then you have a different error
Aitrepreneur
2024-08-14 18:53:31 +0000 UTCI'm having the same exact problem and got the same error and CUDA didn't help with that particular error.
DStiles
2024-08-14 16:05:27 +0000 UTCNo that's not possible no, you can't control that with this repo, I seem to remember that parler tts had this kind of function however, haven't tried it but you can try it out yourself see if that works for you: https://huggingface.co/spaces/parler-tts/parler_tts_mini
Aitrepreneur
2024-06-16 21:10:15 +0000 UTCPlease is there a guide or tip on how to write prompts? For example I want something to be SAID OUT LOUD, or in a certain tone. Is that possible?
Tomas Rampas
2024-06-16 09:44:58 +0000 UTCfirst it's a false positive, antiviruses suck major D.. don't use them ever... Secondly, disable them and run the installer normally, not as admin
Aitrepreneur
2024-06-04 21:05:07 +0000 UTCDo these not work? I get revocation function was unable to check revocation for the certificate on the admin and my antivirus says the ultimate has a trojin virus in it.
Nathan Bryan
2024-06-04 20:21:36 +0000 UTCare you sure that python is installed correctly? If yes then, first open a command prompt window and drag and drop the installer file inside the window, then press enter, this should avoid the window from closing and it will at least give us an error code we can use to troubleshoot the error, make also sure that you have disabled any antivirus/firewall from running and blocking the file/connection.
Aitrepreneur
2024-05-29 12:25:15 +0000 UTCI tried that but now the installer crashes ever time.
YT Super Mario Minecraft
2024-05-28 22:25:12 +0000 UTCYou need to download and install cuda 12 from the nvidia website: https://developer.nvidia.com/cuda-downloads
Aitrepreneur
2024-05-28 19:42:15 +0000 UTCFirst download and install cuda 12 from the nvidia website: https://developer.nvidia.com/cuda-downloads Then that your python version installed is the correct one, you need to uninstall your current python installation and reinstall it correctly. Go to the add and remove programs, search for python and uninstall both the current python version and the python install program. Once this is done, go here and download this installer: https://www.python.org/ftp/python/3.10.6/python-3.10.6-amd64.exe Run it and check the “Add python 3.10 to Path” checkbox and continue with the installation. You can check that the right python version is installed by opening a new command prompt window and typing: python --version and it should give you the 3.10.6 version Then just relaunch the 1-click installer in a new folder and try again.
Aitrepreneur
2024-05-28 19:41:58 +0000 UTCyou can just use the xtts coqui extension or xtts inside sillytavern
Aitrepreneur
2024-05-28 19:41:39 +0000 UTCOn step 1 creating data set, after loading the audio I got this error: The data processing was interrupted due an error !! Please check the console to verify the full error message! Error summary: Traceback (most recent call last): File "C:\Users\walla\Desktop\New folder\xtts-finetune-webui\xtts_demo.py", line 246, in preprocess_dataset train_meta, eval_meta, audio_total_size = format_audio_list(audio_path, whisper_model = whisper_model, target_language=language, out_path=out_path, gradio_progress=progress) File "C:\Users\walla\Desktop\New folder\xtts-finetune-webui\utils\formatter.py", line 131, in format_audio_list segments = list(segments) File "C:\Users\walla\Desktop\New folder\xtts-finetune-webui\venv\lib\site-packages\faster_whisper\transcribe.py", line 1111, in restore_speech_timestamps for segment in segments: File "C:\Users\walla\Desktop\New folder\xtts-finetune-webui\venv\lib\site-packages\faster_whisper\transcribe.py", line 508, in generate_segments encoder_output = self.encode(segment) File "C:\Users\walla\Desktop\New folder\xtts-finetune-webui\venv\lib\site-packages\faster_whisper\transcribe.py", line 767, in encode return self.model.encode(features, to_cpu=to_cpu) RuntimeError: Library cublas64_12.dll is not found or cannot be loaded could you help me please?
Wallace
2024-05-28 13:30:52 +0000 UTCI am geting two errors. This one Using cached gradio-4.13.0-py3-none-any.whl.metadata (15 kB) ERROR: Could not find a version that satisfies the requirement torch==2.1.1 (from versions: 2.2.0, 2.2.1, 2.2.2, 2.3.0) ERROR: No matching distribution found for torch==2.1.1 and this one ERROR: Could not find a version that satisfies the requirement torch==2.1.1+cu118 (from versions: 2.2.0+cu118, 2.2.1+cu118, 2.2.2+cu118, 2.3.0+cu118) ERROR: No matching distribution found for torch==2.1.1+cu118
YT Super Mario Minecraft
2024-05-25 17:13:26 +0000 UTCThis is great! is there any way to get this to work in text-generation-webui with the "automatic xtts to RVC". Maybe an extension for oobabooga already exists to do that? Or do I have to frankenstein the .py files code and try to do it with Chatgpt 4o? Would be great to have a chat with Carl Sagan lol
Karl B
2024-05-23 02:16:29 +0000 UTCopen a command prompt window and type: python -m pip install --upgrade pip Then 1) Open System Properties: Right-click on the Start button and select System. Click on Advanced system settings on the left side. In the System Properties window, click on the Environment Variables button near the bottom. Add New System Variable: In the Environment Variables window, under the "System variables" section, click on New. For the Variable name, enter PYTHONUTF8. For the Variable value, enter 1. Click OK to close the dialog box and exit. 2) Open Control Panel and go to Clock and Region. Click on Region. In the Region window, go to the Administrative tab. Click on the Change system locale button. Ensure that the checkbox for Beta: Use Unicode UTF-8 for worldwide language support is checked. Click OK and then Restart the computer. You might also have a bad initial python install, you need to uninstall your current python installation and reinstall it correctly. Go to the add and remove programs, search for python and uninstall both the current python version and the python install program. Once this is done, go here and download this installer: https://www.python.org/ftp/python/3.10.6/python-3.10.6-amd64.exe Run it and check the “Add python 3.10 to Path” checkbox and continue with the installation. You can check that the right python version is installed by opening a new command prompt window and typing: python --version and it should give you the 3.10.6 version Then just relaunch the 1-click installer in a new folder and try again.
Aitrepreneur
2024-05-18 15:32:18 +0000 UTCwhen trying to install, I am getting this error about the model "requests" Collecting gradio==4.13.0 (from -r .\requirements.txt (line 1)) Using cached gradio-4.13.0-py3-none-any.whl.metadata (15 kB) ERROR: Could not find a version that satisfies the requirement torch==2.1.1 (from versions: 2.2.0, 2.2.1, 2.2.2, 2.3.0) ERROR: No matching distribution found for torch==2.1.1 Looking in indexes: https://download.pytorch.org/whl/cu118 ERROR: Could not find a version that satisfies the requirement torch==2.1.1+cu118 (from versions: 2.2.0+cu118, 2.2.1+cu118, 2.2.2+cu118, 2.3.0+cu118) ERROR: No matching distribution found for torch==2.1.1+cu118 Install deepspeed for windows for python 3.10.x and CUDA 11.8 Traceback (most recent call last): File "D:\aitrepreneur\tts\xtts-webui\scripts\modeldownloader.py", line 4, in import requests ModuleNotFoundError: No module named 'requests' Install complete. Press any key to continue . . . (venv) D:\aitrepreneur\tts\xtts-webui>
yashwanth 439
2024-05-18 00:32:15 +0000 UTCok dm me tomorrow, we could sync up so I can take a look myself on your computer
Aitrepreneur
2024-05-17 22:00:16 +0000 UTCHi again ok but give me more info, do you have the correct python installed now? Yes Have you reinstalled the webui in a new folder again? Yes Do you see any error during the installation? No, it flashes up for a split second and then closes. When I try running via cmd I get: C:\Users\me>C:\Users\me\Downloads\xtts-finetune-dl\xtts-finetune-webui\start.bat 'venv' is not recognized as an internal or external command, operable program or batch file. python: can't open file 'C:\\Users\\me\\xtts_demo.py': [Errno 2] No such file or directory Do you have any antivirus/firewall running? Yes, Norton. The Microsoft firewall I've turned off App control which tends to block stuff. Hope this helps!
James Stewart
2024-05-17 20:47:31 +0000 UTCok but give me more info, do you have the correct python installed now? Have you reinstalled the webui in a new folder again? Do you see any error during the installation? Do you have any antivirus/firewall running?
Aitrepreneur
2024-05-17 10:44:49 +0000 UTCHi K, I did all the steps you said and had exactly the same issue again :(
James Stewart
2024-05-17 06:20:03 +0000 UTCaccent can be finetuned (maybe try with the 2.0.3 model as well might give you better results)
Aitrepreneur
2024-05-16 18:33:01 +0000 UTCI gave this whole process a try and it is pretty amazing just like you mention! One question I have though is if there is any way to preserve accents? I was trying to get a British-sounding voice to work, but it always seems to end up sounding like a more American accent no matter how much I seem to train it. For XTTS, having a voice file with a British accent seems to lose it and for the Finetune as well it doesn't preserve the accent. I had slightly more success with RVC, but half of that was using a voice file that already had an accent. Is there a trick I'm missing or is that just how the model gets trained?
Cory LeGuin
2024-05-16 18:25:30 +0000 UTCI need the full error log, any errors during install also?
Aitrepreneur
2024-05-16 16:18:35 +0000 UTCThis is probably a bad initial python install, you need to uninstall your current python installation and reinstall it correctly. Go to the add and remove programs, search for python and uninstall both the current python version and the python install program. Once this is done, go here and download this installer: https://www.python.org/ftp/python/3.10.6/python-3.10.6-amd64.exe Run it and check the “Add python 3.10 to Path” checkbox and continue with the installation. You can check that the right python version is installed by opening a new command prompt window and typing: python --version and it should give you the 3.10.6 version Then just relaunch the 1-click installer in a new folder and try again. Also no need to write new messages, just reply to this message, or send a dm, easier to follow this way
Aitrepreneur
2024-05-16 16:17:24 +0000 UTCHi finetune xtts is working fine, but when i give it a try about your last video over to open xtts-rvc-ui doesn´t open. thats the output : ValueError: mutable default for field common is not allowed: use default_factory. any idea ?
Emmet Brown
2024-05-16 13:08:13 +0000 UTCThe CMD screen pops up for a split second and then closes/disappears. Not even enough time to see what's written on the screen. I didn't see any errors during install. If dragging into the cmd prompt, the command line is: C:\Users\me>C:\Users\me\Downloads\xtts-finetune-webui\start.bat and the responding message is: 'venv' is not recognized as an internal or external command, operable program or batch file. python: can't open file 'C:\\Users\\me\\xtts_demo.py': [Errno 2] No such file or directory
James Stewart
2024-05-16 08:39:51 +0000 UTCwhat exactly happens? The cmd doesn't even start? Any errors during install? have you tried opening a new cmd prompt and drag and dropping the start bat file inside?
Aitrepreneur
2024-05-15 23:43:20 +0000 UTCI've installed the XTTS-Finetune model as per K's video and the one click installer, however when I go into the folder and click the start.bat file, the program doesn't start/give me a localhost URL. Anyone know what I need to do?
James Stewart
2024-05-15 22:32:22 +0000 UTCopen a command prompt window and type: python -m pip install --upgrade pip Then 1) Open System Properties: Right-click on the Start button and select System. Click on Advanced system settings on the left side. In the System Properties window, click on the Environment Variables button near the bottom. Add New System Variable: In the Environment Variables window, under the "System variables" section, click on New. For the Variable name, enter PYTHONUTF8. For the Variable value, enter 1. Click OK to close the dialog box and exit. 2) Open Control Panel and go to Clock and Region. Click on Region. In the Region window, go to the Administrative tab. Click on the Change system locale button. Ensure that the checkbox for Beta: Use Unicode UTF-8 for worldwide language support is checked. Click OK and then Restart the computer. then redo the installation
Aitrepreneur
2024-05-13 22:47:26 +0000 UTCI keep getting a : Traceback (most recent call last): File "C:\Users\Dayma\Documents\TTs\xtts-webui\app.py", line 1, in from scripts.modeldownloader import install_deepspeed_based_on_python_version File "C:\Users\Dayma\Documents\TTs\xtts-webui\scripts\modeldownloader.py", line 4, in import requests ModuleNotFoundError: No module named 'requests' when trying to run the batch file.
Seth Haynes
2024-05-13 22:24:12 +0000 UTCNo unfortunately i don't know any sources of finetuned xtrs models except huggingface, not sure why tbh maybe there is less demand for it? Compared to rvc models that you can find anywhere
Aitrepreneur
2024-05-13 21:39:37 +0000 UTCOne last question on this if you have the time: I can't find any websites with ready to download voices for this, do you know any? Would be great to start with a large library with already trained voices. Thanks!
Fernando
2024-05-13 10:55:12 +0000 UTCtry this one: https://we.tl/t-JnKo7InjBD
Aitrepreneur
2024-05-12 21:42:26 +0000 UTCmy c drive holds the os and i have all the ai stuff on drive e but i tried installing it on drive c but same error
james
2024-05-12 20:07:36 +0000 UTClooks like you don't have a C drive...on which drive have you installed your OS?
Aitrepreneur
2024-05-12 16:38:46 +0000 UTCim sorry k it seem like i have a issue with everythingDownloading ffmpeg... % Total % Received % Xferd Average Speed Time Time Time Current Dload Upload Total Spent Left Speed 100 282 100 282 0 0 247 0 0:00:01 0:00:01 --:--:-- 248 100 85.9M 100 85.9M 0 0 1727k 0 0:00:50 0:00:50 --:--:-- 3583k Extracting ffmpeg... New-Item : Cannot find drive. A drive with the name 's\Documents\ffmpeg\ffmpeg.zip -DestinationPath C' does not exist. At C:\Windows\system32\WindowsPowerShell\v1.0\Modules\Microsoft.PowerShell.Archive\Microsoft.PowerShell.Archive.psm1:318 char:32 + ... eatedItem = New-Item -Path $DestinationPath -ItemType Directory -Conf ... + ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + CategoryInfo : ObjectNotFound: (s\Documents\ffm...stinationPath C:String) [New-Item], DriveNotFoundExce ption + FullyQualifiedErrorId : DriveNotFound,Microsoft.PowerShell.Commands.NewItemCommand Failed to extract ffmpeg. Please check the zip file and destination path. Press any key to continue . . .
james
2024-05-12 00:27:40 +0000 UTCOoooh how did I miss that one, AWESOME!! Thank you so much!
Fernando
2024-05-11 13:06:39 +0000 UTCyou can try out the 2.0.3 model, this might make it better for some voices, otherwise yeah manually editing it is always good, there are automatic softwares that can do that as well but they are paid
Aitrepreneur
2024-05-11 12:53:36 +0000 UTCRVC, already did a video on this: https://youtu.be/sT6_pi5Jq4w
Aitrepreneur
2024-05-11 12:52:06 +0000 UTCIs there a way to reduce the gaps in the audio output? There are long pauses with the audio that isn't present in the text and I'm trying to figure out how to get rid of them. Or is that something I'll have to do in Audacity?
RustyBaggs
2024-05-11 00:44:22 +0000 UTCGreat video! What would you recommend as the best Github repo for voice conversion (similar to Voicemod) that converts emotion well and has a high quality? In other words, to use a different voice while acting a scene/monologue/et cetera, but with all the nuances that my own voice does. Thanks in advance!
Fernando
2024-05-10 21:33:00 +0000 UTCalready did, this is still xtts, so it's already available inside Sillytavern and text gen webui
Aitrepreneur
2024-05-10 10:07:46 +0000 UTCthe error itself is not an issue, I see this error happening to me as well, but it shouldn't interfere with the use of the tool
Aitrepreneur
2024-05-10 10:06:53 +0000 UTCno there is no API unfortunately
Aitrepreneur
2024-05-10 10:05:27 +0000 UTCwow, this is amazing, can i integrate this with text-generation-webui/SillyTavern? if this is possible, maybe you could make an video about it?
Lars Drakblod
2024-05-10 03:15:31 +0000 UTCrun into this issue with the finetune repo: Exception in callback _ProactorBasePipeTransport._call_connection_lost(None) handle: Traceback (most recent call last): File "C:\Users\MainUser\AppData\Local\Programs\Python\Python310\lib\asyncio\events.py", line 80, in _run self._context.run(self._callback, *self._args) File "C:\Users\MainUser\AppData\Local\Programs\Python\Python310\lib\asyncio\proactor_events.py", line 165, in _call_connection_lost self._sock.shutdown(socket.SHUT_RDWR) ConnectionResetError: [WinError 10054] An existing connection was forcibly closed by the remote host any ideas on how to fix?
Andrew Ostrovsky
2024-05-10 01:13:35 +0000 UTCLooking forward to trying tonight. But, does it allow for an API local server setup so I can presetup voices and then submit text which then does both including RVC, producing a wav/mp3?
Jim Gale
2024-05-09 17:40:55 +0000 UTCAhhh no I want to learn to code to make my own chatbot lol
Carlos.exe
2024-05-09 17:14:37 +0000 UTC
