search

Mark12Ding / FAME

Code for Motion-aware Contrastive Video Representation Learning via Foreground-background Merging (CVPR 2022)
https://arxiv.org/abs/2109.15130
Apache License 2.0
45 stars 5 forks source link

About pre-trained model #2

Open FriedRonaldo opened 2 years ago

FriedRonaldo commented 2 years ago

Hi, thanks for the interesting work!

Could you provide the ckpt files of pre-trained models? I mean R(2+1)D on Kinetics-400, UCF-101 and HMD51, I3D on HMD51.

Thanks!

Mark12Ding commented 2 years ago

Thanks for the attention. The checkpoints will be released when ready.

nanyyyyyy commented 1 year ago

The download link has expired. Would you mind sharing it again? Thank you so much!