Installable package

[Tomiinek]

Hello,

First of all, thank you so much for your excellent work! ❤️

I’m interested in using your models in my projects, and I find it more convenient when things are easily installable and can be imported as Python packages. To make this possible, I made some adjustments to your repository. Now, with a simple command:

pip install git+https://github.com/Tomiinek/WavTokenizer

or from a working directory

pip install -e .

I can use your repository as a package:

import torch
from wavtokenizer import WavTokenizer, convert_audio


wavtokenizer = WavTokenizer.from_pretrained0802(...)

wav = torch.randn(1, 24000)
wav = convert_audio(audio, 24000, 24000, 1) 
bandwidth_id = torch.tensor([0])
with torch.no_grad():
    features, codes = wavtokenizer.encode_infer(wav, bandwidth_id=bandwidth_id)
    audio_out = wavtokenizer.decode(features, bandwidth_id=bandwidth_id) 

I hope you find these changes useful!