Data falsification?

[yuguangs]

Hello Thank you for the Excellent work!

I can't reproduce the results on the paper. Validation on Kitti Dataset is very low. I used YoloStereo3D for the backbone, and changed the files in this repo ( feature_level_generation.py). I used chen_split train_split_file = os.path.join(cfg.path.visualDet3D_path, 'data', 'kitti', 'chen_split', 'train.txt'), val_split_file = os.path.join(cfg.path.visualDet3D_path, 'data', 'kitti', 'chen_split', 'val.txt'),

I think this result is similar to the result without Ldisp,but I didn't change the code

Do you have any suggestions...? Thank you !

[yuguangs]

liga result