/Audio2Face

Implementation with Pytorch but delete the emotional part(to get better result on specific dataset)

Primary LanguagePythonMIT LicenseMIT

Audio2Face

If you are not new to DL, this project does not help to you because the model described in the essay is quite easy to implement.

Maybe update this repo with additional info later, I mean I'll try to make this program easy to run... If you have any urgent demand, please view the works of others or contact me for this repo is now 1000% can not run without modification...

Introduction

2022.07.26

  • model.py described the model implementation(simplified ver)
  • dataset.py CAUTION: this dataset is designed for particular dataset, please replace it when you are using other data
  • train.py very coarse ver od train program, just modify it as you wish
  • infer.py to predict with your trained model