ucf-101 fine-tune from kinetics-400 pretrained model

MCG-NJU / VideoMAE

[NeurIPS 2022 Spotlight] VideoMAE: Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training

https://arxiv.org/abs/2203.12602

Other

1.38k stars 137 forks source link

ucf-101 fine-tune from kinetics-400 pretrained model #71

Open taeoh-kim opened 1 year ago

taeoh-kim commented 1 year ago

Hello.

How do I reproduce K400 (pretrain, 1600ep) -> UCF101 (finetune) performance as shown in the paper (96.1 Top-1 in UCF)? Should I use the same config as UCF101 (pretrain) -> UCF101 (finetune)?

YTHmamba commented 1 year ago

Hello, how did you reproduce it？

YTHmamba commented 1 year ago

hello, @yztongzhan 。Thank you for your contribution. Could you please tell me the configuration of finetune UCF101 with accuracy of 96.1 under K400 (pretrain, 1600ep)