Closed sgzqc closed 1 year ago
Sorry, we lost our pretrained model, but we will release it soon. Using the code here, you can easily implement the indicators in the paper.
@rhwang1314 催一下
+1
if no ckpt, the results likes follow: have on meanings {'f1_score': 0.0, 'precision': 0.0, 'recall': 0.0, 'x_error_close': nan, 'x_error_far': nan, 'z_error_close': nan, 'z_error_far': nan}
hi, all,
You're welcome to contribute your own pre-trained model. Our team's focus has shifted, and we are now analyzing some more interesting conclusions based on the techniques from this paper.
Here are the third-party reproduction results feedback and our suggestions: https://github.com/gigo-team/bev_lane_det/issues/10
Feel free to provide us with your opinions and suggestions.
Thanks.
+1