search

YuanGongND / cav-mae

Code and Pretrained Models for ICLR 2023 Paper "Contrastive Audio-Visual Masked Autoencoder".
BSD 2-Clause "Simplified" License
214 stars 20 forks source link

Could you release the checkpoints pretrained on Kinetics 400 #21

Open qiyue-liang opened 8 months ago

qiyue-liang commented 8 months ago

Hi Yuan, in the paper you mentioned that "Specifically, we train the model on Kinetics-400 (K400) dataset and report the top-1 accuracy on Kinetics-Sounds", I'm wondering if it is possible that you could release the checkpoints pretrained on K400 dataset for action recognition tasks. It appears that the repo currently provides checkpoints only for the event classification tasks. Thanks a lot!

YuanGongND commented 8 months ago

hi there,

Thanks for the inquiry. Unfortunately I do not have time to release this checkpoint. It is not just a checkpoint, but also an inference script (add probably also data preprocessing). Before release, I do a cleanup on my research code and run a brief test. There are many experiments in the appendix and I only have time to release codes for experiments in the main manuscript.

However, I am willing to answer any questions related to that.

-Yuan