pandaset-devkit
pandaset-devkit copied to clipboard
scaleapi
How to evaluating results?
Hi,there I am trying to test kitti trained models like pv_rcnn on pandaset. it looks like there is no standard evaluation metric yet.
can you tell me how to do it？
Hi,there I am trying to test kitti trained models like pv_rcnn on pandaset. it looks like there is no standard evaluation metric yet.
can you tell me how to do it？
I also have this question,how you solve it?
According to the pandaset paper, they are using kitti evaluation metric.
But I don't know how to convert to Kitti format, and the paper doesn't give a tutorial, I'm still stuck in this place, waiting for the official evaluation metric.
By the way, after I converted Pandaset's data into Bin format and tested, there is no result...
Do you have wechat?, I am stuck int this place too I want to talk with you~ my wechat: MegaloAiming
According to the pandaset paper, they are using kitti evaluation metric.
But I don't know how to convert to Kitti format, and the paper doesn't give a tutorial, I'm still stuck in this place, waiting for the official evaluation metric.
By the way, after I converted Pandaset's data into Bin format and tested, there is no result...
Do you get the AP? I reappear pandaset on 3D model ,but I get the higher AP in car ,the lower AP in Ped&cyclist
@Joeless Sorry for the late reply, you can contact me by email. I tried converting pandas to kitti's bin format, but I didn't get any results (AP=0), and then I gave up Debug. BTW, can you share your approach?
@Joeless Sorry for the late reply, you can contact me by email. I tried converting pandas to kitti's bin format, but I didn't get any results (AP=0), and then I gave up Debug. BTW, can you share your approach?
I just transfrom to kitti's format,using kitti eval