- Python=3.6
- Pytorch=1.6.0
FAIR-Play can be accessed here. YT-ASMR can be accessed here.
-
Prepare datasets. Please prepare the dataset as the instructions in FAIR-Play.
-
Training.
./train.sh
- Ealuation.
./test.sh
A set of pretrained weights can be found at https://drive.google.com/drive/folders/1N7UMOZqNbFe_QXx4x_kKH02CQI8_DwPa?usp=sharing .
We borrowed a lot of code from https://github.com/SheldonTsui/SepStereo_ECCV2020 and https://github.com/facebookresearch/2.5D-Visual-Sound. Thanks for their great works. Please also cite their nice works if you use this code.