about the structure of sal_unet

[watsy1314]

Hello Author! I would like to ask, the source code both visual.py and audio-visual.py call sal_unet in saliency_decoder, which doesn't match the structure proposed in the paper, ah? What's going on here?

[junwenxiong]

Can you provide the details?