train or val ,pretrained models?

facebookresearch / Listen-to-Look

Listen to Look: Action Recognition by Previewing Audio (CVPR 2020)

Creative Commons Attribution 4.0 International

126 stars 15 forks source link

train or val ,pretrained models? #3

Closed ctx2333-ctx closed 3 years ago

ctx2333-ctx commented 3 years ago

Dear Doctor: The features you provided is fussy to distinguish whether it‘train set or val set.I beg for the clearly divided dataset（audio+visual）or features. What's more ,I can't find the pretrained model. Thanks！

rhgao commented 3 years ago

The split and the ImageAudio model for ActivityNet is shared here: https://utexas.box.com/s/8rxijymjo0kx4e9jk7byobdc8vnf377l

ctx2333-ctx commented 3 years ago

The split and the ImageAudio model for ActivityNet is shared here: https://utexas.box.com/s/8rxijymjo0kx4e9jk7byobdc8vnf377l

get it ,thanks!