search

noahzn / Lite-Mono

[CVPR2023] Lite-Mono: A Lightweight CNN and Transformer Architecture for Self-Supervised Monocular Depth Estimation
MIT License
540 stars 61 forks source link

Generalization results of litemono on make3d #73

Closed bizheyandebing closed 1 year ago

bizheyandebing commented 1 year ago

This is a very valuable job. I have some confusion, I can reproduce your results in Table 2 based on the model you provided,

image

but the results are different from Table 3 when I use the same model for make3d generalization testing. image

According to the description in the article, they should be the same model. Is there something wrong with my evaluation code?

bizheyandebing commented 1 year ago

I think I solved this problem, you answered me before in monodepth2's discussion forum, but I forgot. Thank you very much, your work is very valuable.

https://github.com/nianticlabs/monodepth2/issues/392#issuecomment-1128977268