ERROR install requirements

[Abbadon999]

I can't install your bot, I'm stuck on the first step of the installation First pip install -t ./requirements.txt gives an error - ERROR: Target path exists but is not a directory, will not continue.

Secondly, I would like more detailed instructions about building as I honestly don't know what to do next. For example is it necessary to insert API or maybe to change input language?

[HRNPH]

oh sorry it's -r flag not -t flag

pip install -r ./requirements.txt

[HRNPH]

Question1

I've already updated the typo in README b81150735366ab8f94af98f526c41c4cfc580169 Try following it again if it worked (Please notify me if it does work) so I can close this Issue with ease

Question2

the instruction available in README includes all the steps necessary for the system to run properly so just follow it step by step without skipping the order (You'll see a blue text leading to the Installation page for the external program)

you don't need to change any default config or provide any API key(We don't use external AI Models) Currently Supported Language was only in English for "Chatting" and Japanese for "Waifu Sound" (AKA you can't choose) This will be implemented for more customization in the near future👍

[Abbadon999]

Okay, thanks for the explanation, but I still can't install the libraries, because I can't install pyopenjtalk-0.3.0. About the -r flag, I forgot to write that I already tried it and it gave me an error just because of pyopenjtalk-0.3.0. I also tried installing pyopenjtalk via pip install --no-binary :all: pyopenjtalk==0.3.0 using pip install cython numpy setuptools wheel but that doesn't help...

[Adolfiux]

You need to install Visual Studio and Visual Studio Build Tools. The trick is to select the C++ tools during the installation of both. After that, you need to open a CMake project in Visual Studio so that the project information displays the location of cmake.exe. Then, you need to add this location to the system's path without including the actual cmake.exe file, only the path where it is located. Once you do this, you should be able to install that and other packages without any issues.

[HRNPH]

You need to install Visual Studio and Visual Studio Build Tools. The trick is to select the C++ tools during the installation of both. After that, you need to open a CMake project in Visual Studio so that the project information displays the location of cmake.exe. Then, you need to add this location to the system's path without including the actual cmake.exe file, only the path where it is located. Once you do this, you should be able to install that and other packages without any issues.

0h yeah I forgot to include that thx!

[HRNPH]

@Abbadon999

Okay, thanks for the explanation, but I still can't install the libraries, because I can't install pyopenjtalk-0.3.0. About the -r flag, I forgot to write that I already tried it and it gave me an error just because of pyopenjtalk-0.3.0. I also tried installing pyopenjtalk via pip install --no-binary :all: pyopenjtalk==0.3.0 using pip install cython numpy setuptools wheel but that doesn't help...

Can you capture the screenshot of an error?

[ponmak]

Collecting pyopenjtalk==0.3.0 Using cached pyopenjtalk-0.3.0.tar.gz (1.5 MB) Installing build dependencies ... done Getting requirements to build wheel ... error error: subprocess-exited-with-error

× Getting requirements to build wheel did not run successfully. │ exit code: 1 ╰─> [23 lines of output]

note: This error originates from a subprocess, and is likely not a problem with pip. error: subprocess-exited-with-error

I have a same issue, Here a some screenshot.

[Abbadon999]

@Adolfiux I already have Visual Studio installed, is there any way I can install C++ and all the features you mentioned earlier into it or do I need to reinstall it?

[HRNPH]

@Abbadon999 @ponmak It appear that you didn't have Cmake (used to build C/C++ program Open-Jtalk in this case) first off all try check if you have Cmake install

# open CMD or Terminal and type
cmake #(you wouldn't have it but just to be sure)

Then referring to this issue in pyopenjtalk you need to install Cmake just download the installer and set it up

hope this help, Please notify me if it work or it didn't 👍

[Abbadon999]

I believe I have Сmake installed

[HRNPH]

@Abbadon999 what about these guys? do you have it installed

gcc (g++)
cmake # you have it
cython

[ponmak]

Thx I got it

[HRNPH]

It seems that @ponmak error and @Abbadon999 error were different

so can I request @Abbadon999 to reproduce the error and send us the error log?

[ponmak]

@Abbadon999

I already have Visual Studio installed, is there any way I can install C++ and all the features you mentioned earlier into it or do I need to reinstall it?

1. Open your Visual Studio Installer. Go to Modify 2.Click to desktop development with c++ and install

Maybe it helped

[Abbadon999]

Hooray, I finally installed "pyopenjtalk-0.3.0", but now I have a new error with "numpy", I guess I should try to change the version, because apparently it conflicts with other packages.

[HRNPH]

seem like you need to downgrade to 1.23.x or 1.20.3 @Abbadon999 if what version work I'll change the numpy requirement to it so please also notify me

[Abbadon999]

I seem to be able to install all the packages(And I was only able to install packages with the version of numpy - 1.23.5), but the python ./main.py gives this error:

Initializing... Dependencies
Traceback (most recent call last):
  File "F:\AI-Waify\NICE\AIwaifu\main.py", line 3, in <module>
    from anime_tts.custom_inference import tts # text to speech from huggingface
  File "F:\AI-Waify\NICE\AIwaifu\anime_tts\custom_inference.py", line 7, in <module>
    net_g_ms, hps = infer.load_model(config_json, pth_path)
  File "F:\AI-Waify\NICE\AIwaifu\anime_tts\infer.py", line 61, in load_model
    _ = utils.load_checkpoint(pth_path, net_g_ms)
  File "F:\AI-Waify\NICE\AIwaifu\anime_tts\utils.py", line 41, in load_checkpoint
    checkpoint_dict = load(checkpoint_path, map_location='cpu')
  File "C:\Users\valer\AppData\Local\Programs\Python\Python310\lib\site-packages\torch\serialization.py", line 771, in load
    with _open_file_like(f, 'rb') as opened_file:
  File "C:\Users\valer\AppData\Local\Programs\Python\Python310\lib\site-packages\torch\serialization.py", line 270, in _open_file_like
    return _open_file(name_or_buffer, mode)
  File "C:\Users\valer\AppData\Local\Programs\Python\Python310\lib\site-packages\torch\serialization.py", line 251, in __init__
    super(_open_file, self).__init__(open(name, mode))
FileNotFoundError: [Errno 2] No such file or directory: './anime_tts/1374_epochs.pth'

[HRNPH]

okay this is due to the model weight inside the repo which needs to be pulled down with GIT LFS or you can download it manually HERE and place in the error location

in the future we'll place all the weight in hugging face and do the download automatically so this wouldn't happen :DD

[Abbadon999]

Edited: So, I reinstalled the files and "1374_epochs.pth" seems to be found, but now a new error....

PS F:\AI-Waify\NICE\AIwaifu> python ./main.py
Initializing... Dependencies
Traceback (most recent call last):
  File "F:\AI-Waify\NICE\AIwaifu\main.py", line 3, in <module>
    from anime_tts.custom_inference import tts # text to speech from huggingface
  File "F:\AI-Waify\NICE\AIwaifu\anime_tts\custom_inference.py", line 7, in <module>
    net_g_ms, hps = infer.load_model(config_json, pth_path)
  File "F:\AI-Waify\NICE\AIwaifu\anime_tts\infer.py", line 61, in load_model
    _ = utils.load_checkpoint(pth_path, net_g_ms)
  File "F:\AI-Waify\NICE\AIwaifu\anime_tts\utils.py", line 41, in load_checkpoint
    checkpoint_dict = load(checkpoint_path, map_location='cpu')
  File "C:\Users\valer\AppData\Local\Programs\Python\Python310\lib\site-packages\torch\serialization.py", line 795, in load
    return _legacy_load(opened_file, map_location, pickle_module, **pickle_load_args)
  File "C:\Users\valer\AppData\Local\Programs\Python\Python310\lib\site-packages\torch\serialization.py", line 1002, in _legacy_load
    magic_number = pickle_module.load(f, **pickle_load_args)
_pickle.UnpicklingError: invalid load key, 'v'.

[HRNPH]

did the file corrupt? how did you "reinstall" if there are no git lfs or you didn't download directly from hugginface there's a chance that the file is just a pointer file and it is not an actual model weight so first of all check the file size it should be around 159 MB

[HRNPH]

@Abbadon999 is it Working?

[Abbadon999]

I reinstalled everything and it all worked, but now it gives an error on startup:

F:\AI-Waify\NICE\AIwaifu>python ./main.py
Initializing... Dependencies
Initializing... Vtube Studio
Error:  [WinError 10061] Connection not established because the destination computer rejected the connection request
This usually means that VTube Studio is not running or the public API is not enabled.
The permissions are not set correctly and/or got rejected.
Traceback (most recent call last):
  File "F:\AI-Waify\NICE\AIwaifu\vtube_studio.py", line 45, in send
    self.websocket.connect(f"ws://localhost:{self.port}")
  File "C:\Users\valer\AppData\Local\Programs\Python\Python310\lib\site-packages\websocket\_core.py", line 249, in connect
    self.sock, addrs = connect(url, self.sock_opt, proxy_info(**options),
  File "C:\Users\valer\AppData\Local\Programs\Python\Python310\lib\site-packages\websocket\_http.py", line 130, in connect
    sock = _open_socket(addrinfo_list, options.sockopt, options.timeout)
  File "C:\Users\valer\AppData\Local\Programs\Python\Python310\lib\site-packages\websocket\_http.py", line 205, in _open_socket
    raise err
  File "C:\Users\valer\AppData\Local\Programs\Python\Python310\lib\site-packages\websocket\_http.py", line 185, in _open_socket
    sock.connect(address)
ConnectionRefusedError: [WinError 10061] Connection not established because the destination computer rejected the connection request

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "F:\AI-Waify\NICE\AIwaifu\main.py", line 16, in <module>
    waifu = Char_control(port=8001, plugin_name='MyBitchIsAI', plugin_developer='HRNPH')
  File "F:\AI-Waify\NICE\AIwaifu\vtube_studio.py", line 89, in __init__
    super().__init__(plugin_name, plugin_developer, port)
  File "F:\AI-Waify\NICE\AIwaifu\vtube_studio.py", line 23, in __init__
    auth_status = self.auth()
  File "F:\AI-Waify\NICE\AIwaifu\vtube_studio.py", line 34, in auth
    result = self.send("AuthenticationTokenRequest", msg)
  File "F:\AI-Waify\NICE\AIwaifu\vtube_studio.py", line 60, in send
    raise Exception("Please start VTube Studio and enable the public API in the settings.")
Exception: Please start VTube Studio and enable the public API in the settings.

[HRNPH]

well Read an exception "Exception: Please start VTube Studio and enable the public API in the settings" basically, just follow the next instructions before starting the script you need to install Vtube Studio

[Abbadon999]

Yay, it works, but I have a question about the sound - How do I make the model open its mouth when the answer is voiced? I tried using the VB-Cable, but then I can't hear anything that's going on in the system... How did you do it?

[HRNPH]

just follow instruction given in the readme you need to install plugin by lua lucky

[Abbadon999]

Thanks for your help, I think this can be closed)

[HRNPH]

There'll be update soon in this months :DD

[Sylensar]

@HRNPH Just got Cmake.

[HRNPH]

@HRNPH Just got Cmake.

if you're on windows don't forget to install C++ Dev SDK

after that proceed with the installation

[Sylensar]

@HRNPH im running the

pip install pyopenjtalk

command correct?

[HRNPH]

@HRNPH im running the

pip install pyopenjtalk

command correct?

yeah