Limin Wang

You need to check whether your file name and file path is correct.

The sever breaks down. We are fixing it. Please stay tuned!

Our paper on MixViT is under preparation. Please stay tuned.

Hi @NielsRogge , Thanks for your update. We have created an organization account on the hub: https://huggingface.co/MCG-NJU You can use this organization for storing our model checkpoints. BTW, you could...

@NielsRogge Any update？

@NielsRogge , Thanks a lot for your help!

Your understanding is right. The normalized pixel loss is from the original image mae paper. If you want to use videomae for reconstruction, you could refer to the original image...

Thanks for your interest in our dataset. Currently, we do not plan to organize any competition in the nearby conference. You could submit your result to evaluation server in codalab.

https://github.com/wanglimin/improved_trajectory/blob/master/src/DenseTrackStab.cpp This is the code to extract trajectories from video.