For Advanced users

　PVTLeaf allows users to implement their own server-client model with any translation engine for experimental purposes. Scripts for experimental servers are available. 

GitHub:

Environment using Google Translate for text translation

Faster-Whisper GoogleTrans Server (Light weight & high ~ middle precision)

The following configurations can operate completely under local environment.

For those who value confidentiality.

Faster-Whisper_Llama3_Server (Very heavy load, Not real time)

Faster-Whisper_M2M100_Sever (Light weight & middle precision)

 

Note: If you do not want to change the version of CUDA or if you have trouble installing it properly, we have packaged the necessary DLLs  in the URL below. Extract it to the same location as the server code above.

　

     CUDA library for windows 

 

If you want to handle a higher load, you can consider using Gunicorn+gevent worker, Kubernetes, etc. to scale the configuration. The client configuration and a brief description are provided in the file below.

 

C:\Users\<UserName>\AppData\Local\omicronware\PVTleaf\user_env\environment.json

 

 

Simple usage of the client (as an experimental server/client) :

Before starting the client PetitVoiceTranslator_Leaf, edit the json file.

 

         "active": "True"

 

 You can connect to a remote server: (e.g. sample.server.com) via HTTP. If you use HTTPS, please prepare a separate certificate and private key on the server side and put them in the same directory as the script. Also, send the server certificate to the client side and edit the json file. The following is an example including the location. Server certificate: (e.g. server.crt)

 

         "server_url": "https://sample.server.com:9443/transcribe"
         "ssl_verify": "C:/Users/<UserName>/AppData/Local/omicronware/PVTleaf/user_env/server.crt",

 

Also, although not recommended, if the client does not need the server's certificate,

 

         "ssl_verify": "False",

 
That's it, thanks.