/voice-vector-web

Primary LanguagePython

Voice Vector Web

How to set up?

Server Side (Tensorflow serving)

python 2.7

Installation

https://www.tensorflow.org/serving/setup

Export model

  • python deploy/export_model.py [case] [path]
    • Ex) python deploy/export_model.py voxceleb_large_4s /data/private/voice-vector/saved_model/

Run tensorflow model server

nohup tensorflow_model_server --port=1027 --model_name=voice_vector --model_base_path=/data/private/voice-vector/saved_model/ &

Client Side (Flask)

python 3.5

Install

ffmpeg

sudo apt-get install ffmpeg

tensorflow serving api

pip install tensorflow-serving-api-python3

https settings

refer to https://blog.miguelgrinberg.com/post/running-your-flask-application-over-https

Run server

python run.py