Strange size of output files for train

aim-uofa / AdelaiDepth

This repo contains the projects: 'Virtual Normal', 'DiverseDepth', and '3D Scene Shape'. They aim to solve the monocular depth estimation, 3D scene reconstruction from single image problems.

Creative Commons Zero v1.0 Universal

1.06k stars 144 forks source link

Strange size of output files for train #54

Closed ori30ffs closed 1 year ago

ori30ffs commented 1 year ago

Hello, I tried to use your train_demo.sh script. After 200 epoch I don't have any progress on test dataset. I attached files with my environment, log and output folder. Could you help me, what I'm doing wrong? Oct11-14-37-36_MSISPlap03.txt Untitled env.txt

guangkaixu commented 1 year ago

The train_demo.sh is only performed to ensure the accurate environment installation, which is only trained on 5 RGB-D pairs. To validate more frequently, you can change the parameter val_step to smaller.