Much lower loss but wrong output

[yang-ruixin]

Hi authors,

Thanks for your great work.

I did an experiment:

1. I trained a model successfully, with loss around 0.6, and the output were meaningful.
2. I did gamma correction for all the input images.
3. I trained the model again, got a much lower loss only aound 0.2, but the output were wrong.

https://github.com/elliottwu/unsup3d/assets/39927800/d8c7cff0-3768-407a-977f-a7225b8ef6ce

https://github.com/elliottwu/unsup3d/assets/39927800/34f1f667-d3e8-42bb-993c-d9fb9cd8e83d

Any suggestions/explanations about lower loss but wrong output?

Thanks in advance.

[yang-ruixin]

The two videos can't show above, so I add two links, thanks.

链接：https://pan.baidu.com/s/1IUQuDMEFjk4-W3xLCg-tww?pwd=l8xi 提取码：l8xi --来自百度网盘超级会员V4的分享

链接：https://pan.baidu.com/s/1PlxL7eOZYB9yYZmId_aBRw?pwd=ngep 提取码：ngep --来自百度网盘超级会员V4的分享