BEVStereo
BEVStereo copied to clipboard
Megvii-BaseDetection
Official code for BEVStereo
# Welcome update to OpenMMLab 2.0 I am Vansin, the technical operator of OpenMMLab. In September of last year, we announced the release of OpenMMLab 2.0 at the World Artificial...
Hi, thanks for opening access. By the way, where should I modify if I want to train with 1408x512 images? I only modified the "final_dims" and the results seem to...
Hello, You have uploaded trained weights in benchmark section. Are they trained with the full nuScenes dataset?
Hi, I'm trying to reproduce the provided results on my cluster. I found the training time is really slow, and the depth loss seems unstable. I'm working on checking the...
Hi! I have a question about the lidar data. To train bevstereo, lidar depth information is necessary? And without lidar information in the training process, how the performance of this...
您好，我在将bacobone替换成vovnet(v2-99)的过程中，遇到了上述问题，具体报错信息如下: Traceback (most recent call last): File "/home/code/BEVStereo-master/exps/bev_stereo_lss_r50_256x704_128x128_20e_cbgs_2key_da_ema.py", line 90, in run_cli() File "/home/code/BEVStereo-master/exps/bev_stereo_lss_r50_256x704_128x128_20e_cbgs_2key_da_ema.py", line 86, in run_cli main(args) File "/home/code/BEVStereo-master/exps/bev_stereo_lss_r50_256x704_128x128_20e_cbgs_2key_da_ema.py", line 56, in main trainer.fit(model) File "/opt/conda/lib/python3.7/site-packages/pytorch_lightning/trainer/trainer.py", line 772,...
Acctually, the depth of foreground should vary smoothly, however, during downsample, lidar depths vary greatly between adjacent areas. 
Hi, I confused about why do the key sensor to sweep sensor operation again in the line [116](https://github.com/Megvii-BaseDetection/BEVStereo/blob/62711e779472dac5907b2622b227416e3694e131/layers/backbones/lss_fpn.py#L1116) 
Hi, I'm confused about the role of the self.use_mask in the line [748](https://github.com/Megvii-BaseDetection/BEVStereo/blob/62711e779472dac5907b2622b227416e3694e131/layers/backbones/lss_fpn.py#L748) I can't find any explanation of it in the paper.
Hi, I'm confused about the two kinds of ego2global value in the line [342](https://github.com/Megvii-BaseDetection/BEVStereo/blob/master/dataset/nusc_mv_det_dataset.py#:~:text=%23%20sweep%20ego%20to%20global) and in the line [419.](https://github.com/Megvii-BaseDetection/BEVStereo/blob/master/dataset/nusc_mv_det_dataset.py#:~:text=mean%20pose%20of%20all%20cams.) Why there are two kinds of ego2global. What's the difference between...