Question about training

snap-research / articulated-animation

Code for Motion Representations for Articulated Animation paper

https://snap-research.github.io/articulated-animation/

Other

1.21k stars 348 forks source link

Question about training #28

Open chaerlo opened 3 years ago

chaerlo commented 3 years ago

Hi, great work! I have some questions about the training part. 1、Are the models in your provided checkpoints trained by the mode of AVD training(train_avd)? 2、Dose bg_predictor not work when AVD training?

AliaksandrSiarohin commented 3 years ago

All provided checkpoints are trained for both stages. Bg is not needed for avd mode.