search

YudongGuo / AD-NeRF

This repository contains a PyTorch implementation of "AD-NeRF: Audio Driven Neural Radiance Fields for Talking Head Synthesis".
MIT License
1.04k stars 178 forks source link

replacing the deepspeech code #125

Open laurentbenaroya opened 1 year ago

laurentbenaroya commented 1 year ago

Hello, I don't want to use deepspeech to extract audio features because it's not compatible with tensorflow 2. Ideally, I would like to work with librosa as it is a reference in audio processing. Could you give me some advice on how to convert the part of your code that uses a deepseech network to another package? Thanks in advance

laurentbenaroya commented 1 year ago

I found an ad-hoc solution, create a second conda environment with conda create -n tf15 python tensorflow=1.15 for the audio analysis