Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Different results on MMdetection3D and OpenPCDet for DAIR-V2X-I training #80

Open
swwwu opened this issue Dec 18, 2023 · 1 comment
Open
Labels
question Further information is requested

Comments

@swwwu
Copy link

swwwu commented Dec 18, 2023

Hi there. I have a question about the evaluation results. I am trying to train PointPillars based on the DAIR-V2X-I dataset with MMdetection3D and OpenPCDet. I got results similar to those of the authors of DAIR-V2X while testing on MMdetection3D. (My results on MMdetection3D: Car-3D AP(R11-iou0.5): 63.58, 54.49, 54.49, DAIR-V2X paper: 63.07, 54.00, 54.01.) However, the result on OpenPCDet is much higher than that on MMdetetion3D. (My results on OpenPCDet: Car-3D AP(R11-iou0.5): 96.67, 90.05, 90.05.) I am very confused about this. Hope someone can help me find out the answer. Thank you very much.

Result on MMdetection3D:
----------- AP11 Results ------------
Car [email protected], 0.70, 0.70:
bbox AP11:24.1993, 19.3548, 19.5493
bev AP11:63.4588, 54.3793, 54.3846
3d AP11:63.0491, 53.9908, 53.9994
aos AP11:12.69, 10.15, 10.25
Car [email protected], 0.50, 0.50:
bbox AP11:24.1993, 19.3548, 19.5493
bev AP11:63.5912, 54.5022, 54.5028
3d AP11:63.5819, 54.4903, 54.4911
aos AP11:12.69, 10.15, 10.25
----------- AP40 Results ------------
Car [email protected], 0.70, 0.70:
bbox AP40:22.6050, 18.6080, 18.7952
bev AP40:67.2324, 54.7689, 54.7698
3d AP40:66.6725, 51.8965, 51.8953
aos AP40:11.85, 9.75, 9.85
Car [email protected], 0.50, 0.50:
bbox AP40:22.6050, 18.6080, 18.7952
bev AP40:67.4173, 54.9247, 54.9263
3d AP40:67.3984, 54.9069, 54.9088
aos AP40:11.85, 9.75, 9.85

Result on OpenPCDet:

Car [email protected], 0.70, 0.70:
bbox AP:27.9997, 24.0742, 24.2758
bev AP:96.2945, 89.6361, 89.6418
3d AP:86.9864, 78.0029, 78.0308
aos AP:14.65, 12.59, 12.70
Car [email protected], 0.70, 0.70:
bbox AP:27.4314, 23.6478, 23.8782
bev AP:97.8024, 92.4339, 92.4493
3d AP:89.8588, 81.2604, 81.2895
aos AP:14.35, 12.37, 12.49
Car [email protected], 0.50, 0.50:
bbox AP:27.9997, 24.0742, 24.2758
bev AP:97.0797, 90.5134, 90.5106
3d AP:96.6720, 90.0508, 90.0580
aos AP:14.65, 12.59, 12.70
Car [email protected], 0.50, 0.50:
bbox AP:27.4314, 23.6478, 23.8782
bev AP:98.8514, 95.7512, 95.7589
3d AP:98.4157, 94.9547, 94.9802
aos AP:14.35, 12.37, 12.49

@haibao-yu haibao-yu added the question Further information is requested label Jan 12, 2024
@bbzh
Copy link

bbzh commented Aug 20, 2024

Hi @swwwu, did you figure out this? Can you share your experiences?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
question Further information is requested
Projects
None yet
Development

No branches or pull requests

3 participants