Closed PlutoXN closed 1 year ago
Is it possible to convert the input camera into a map in real time like yolo?
MapTR already demonstrated much higher efficiency compared to prior arts. For even higher efficiency like YOLO (note that YOLO is used on single-camera setting), you can reduce the size of input 6 camera images.
Do you need to train it every time before visualizing it?
No, we don't. The trained models are ready for direct applications just like YOLO.
How to reproduce what is in this video of End-to-end Planning?
Please refer to VAD
Is it possible to delete the number of cameras?
Yes, MapTR is flexible to the number of cameras.
Hello, is it possible to convert the input camera into a map in real time like yolo? Do you need to train it every time before visualizing it? How to reproduce what is in this video of End-to-end Planning? Is it possible to delete the number of cameras?