wancheng Shen

How can I generate these tfrecord files myself? So I don't need to download it...

Cross Attention

3

I am new to lingvo and I would like to know where to find the code for Cross Attention in `DeepFusion: Lidar-Camera Deep Fusion for Multi-Modal 3D Object Detection`?

我的pcd文件是读取雷达数据然后转成的，每一帧点的数量是24000，每次启动时加载pcd文件比较慢，标注完一帧数据后，切换到下一帧，耗时也比较多，不知道这是正常的情况吗？

Is there any ablation exps about the flag `task_specific`, the gerneral method is weightet_sum https://github.com/HuangJunJie2017/BEVDet/blob/dev3.0/mmdet3d/models/dense_heads/centerpoint_head.py#L293

onnx file

1

Hi, shall you update the onnx file address? the old address was invalid....

https://github.com/Yzichen/FlashOCC/blob/master/projects/mmdet3d_plugin/models/dense_heads/bev_occ_head.py#L286 Is `softmax` necessary when we get semantic occupancy results from the outputs of the model?

Great work! Compared with obj-flow,occ-flow can provide more accurate information about the scene. Currently, i already have a occ3d dataset, how can i get the related occ-flow?