OpenDriveLab / maskalign

[CVPR 2023] Official repository for paper "Stare at What You See: Masked Image Modeling without Reconstruction"
Apache License 2.0
62 stars 5 forks source link

can you provide the loss log #3

Closed valencebond closed 9 months ago

valencebond commented 1 year ago

I use your code to pretrain video datasets, But I'm not sure what is the normal rate of loss decline

HellwayXue commented 1 year ago

Hi, due to the internship change, I have no access to the previous log. But reproducing Maskalign is very easy, so my suggestion is first pre-training on ImageNet and then checking the loss status. You can also discuss more details with me.