search

HKUST-Aerial-Robotics / Stereo-RCNN

Code for 'Stereo R-CNN based 3D Object Detection for Autonomous Driving' (CVPR 2019)
MIT License
690 stars 177 forks source link

Can not reproduce the reported results #7

Closed zklgame closed 5 years ago

zklgame commented 5 years ago

I have tried several times to reproduce the reported results in the Paper using the public code and config setting, but failed each time. I got results that were 7% worse than the reported number when I used train.txt for training and val.txt for validation.

Then I decided to run an interesting experiment that I used trainval.txt for training and val.txt for validation, I can reproduce the Paper number!

Followings are my trained models on train.val and trainlog.txt:

https://pan.baidu.com/s/1zhCt1P3KsVrCyBk8I8kkew https://pan.baidu.com/s/1P7IXetc1otXtDf0HJQ_7Ug

If any question, please discuss freely.

Jokoe66 commented 5 years ago

is result evaluated on val set reported in the paper?

Jokoe66 commented 5 years ago

btw, could you please share your trained models after each epoch? I'd like to check my training process.