Python Audio Transcription for Mandarin

• Linux / Mac OS
• Windows OS, run it using Docker
• Python, version 3.7 - 3.11
• Tensorflow, version 0.37.0 (use in this project)
• Tensorflow IO, version 2.16.2 (use in this project)
• librosa, version 0.10.2.post1 (use in this project)

Linux and Mac OS

Go to project directory

cd path/to/python/project/dir/py-audio-transcription-zh

Install the required library

pip install tensorflow==2.16.2 tensorflow-io==0.37.0

Using Docker

Run it with this Docker Image that contains required library :

docker pull krlan2789/python-tensorflow:1.0.1
docker run -it --rm -v path/on/your/machine/dir/py-audio-transcription-zh:path/on/docker/container/py-audio-trancription-zh krlan2789/python-tensorflow:1.0.1 bash

# Example
docker run -it --rm -v D:/Files/Documents/Python/Projects/py-audio-transcription-zh:/tmp/py-audio-trancription-zh krlan2789/python-tensorflow:1.0.1 bash

After entering the Docker Container terminal, run command below :

cd py-audio-trancription-zh

Scripts

Create Dataset

Run main.py to creating dataset.json file :

python main.py

Train Model

Run train.py to train model and save to .h5 file :

python train.py

Test Model

Run transcribe.py for testing the model :

python transcribe.py