search

pqhieu / jsis3d

[CVPR'19] JSIS3D: Joint Semantic-Instance Segmentation of 3D Point Clouds
https://pqhieu.com/research/cvpr19/
MIT License
175 stars 36 forks source link

Cannot reproduce paper results #7

Closed zhangguanghui1 closed 5 years ago

zhangguanghui1 commented 5 years ago

I run your code according to the current Readme file with MV-CRF. I use the same configuration present in s3dis.json but I cannot obtain the same results evidenced in the paper. The platform: python 3.5.2, torch 1.0, cuda 9.0 cudnn 7.3.0 After 100 epoch,

Overall accuracy: 0.829 Mean accuracy: 0.507 Mean IoU: 0.433 Mean precision: 0.389 Mean recall: 0.287

Can you help me? Thank you very much.

pqhieu commented 5 years ago

Hi @zhangguanghui1,

Sorry for the late reply. I was away so I couldn't look into your issue. As I said before in #3, you can get results that are similar to the paper if you log an earlier model (around 40 epochs). I have changed the default training config accordingly to clear any confusion.

HaiyongJiang commented 4 years ago

Hi @zhangguanghui1, do you resolve the problem? I also got similar performance as yours. Do you have any suggestion to improve it? Thank you.