sasa_nuscenes model

[liuhuaijjin]

Dear author, can you provide the nuscenes pretrained model

[pummi823]

hello, have you run the code on nuscenes dataset? I run it by batch_size=12, and other parameters are the same, but only get map around 20......

[twn29004]

hello, have you run the code on nuscenes dataset? I run it by batch_size=12, and other parameters are the same, but only get map around 20......

Hello, have you reproduced the results in the paper?

[pummi823]

Hello, for KITTI, yes, for nuScenes, no T T

[ihaohe]

I might reproduce the result. The best performace is as blow. But still a litter lower than the paper declared. Just use the cfg tools/cfgs/nuscenes_models/3dssd_sasa.yaml. It seems the batchsize matters.