MCG-NJU SparseBEV issues

MCG-NJU / SparseBEV

[ICCV 2023] SparseBEV: High-Performance Sparse 3D Object Detection from Multi-Camera Videos

https://arxiv.org/abs/2308.09244

MIT License

305 stars 20 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

尺度自适应自我注意力机制

#79 1yuechu1 opened 6 days ago
1
pkl问题

#78 lht6806 closed 2 days ago
5
训练出现nan 怎么解决

#77 rubbish001 opened 1 week ago
0
cuda11.8

#76 ChenJian7578 closed 2 days ago
1
关于calculate absolute velocity according to time difference

#75 kingofstu opened 2 weeks ago
1
关于sparsebev_sampling中的sample_points_cam以及5-D tensor采样问题

#74 kingofstu opened 3 weeks ago
1
关于self.num_points和self.num_levels

#73 kingofstu closed 3 weeks ago
2
如何生成nuscenes_info_train.pkl文件？

#72 kingofstu closed 1 month ago
4
cu文件用什么编译器编写的，怎么调试啊

#71 rubbish001 opened 1 month ago
0
What's the _delete_ parameter means?

#70 ZitengXue opened 1 month ago
1
decoder layer里的add & norm

#69 Azeral0619 opened 1 month ago
1
牛，训练11天，还是a100，普通人还能跑吗

#68 rubbish001 closed 1 month ago
0
pretraining SparseBEV with only 1 frame

#67 jeff770 opened 2 months ago
1
sampling_4d

#66 jhjun724 closed 2 months ago
2
Performance of sparse voxel decoder

#65 Yzichen closed 3 months ago
3
sampling points　lidar2img问题

#64 zhaoyangwei123 closed 3 months ago
3
About max points in msmv_sampling.cpp

#63 zbaishancha opened 3 months ago
3
OverflowError: cannot convert float infinity to integer

#62 zhaoyangwei123 closed 3 months ago
3
SparseOcc当中sparse decoder部分的最后一层query数量是65536吗？以及第一部分的体素化分辨率是1.6m*1.6m*1.6m吗？

#61 cheng-peng-git closed 2 months ago
8
self.init_query_bbox = nn.Embedding(self.num_query, 10) . This parameter doesn't seem to have been updated.

#60 zbaishancha closed 3 months ago
6
Why is there a problem with tensorboard's events file?

#59 zbaishancha closed 3 months ago
1
请问作者有没有做过不同query数量以及不同数量尺度特征图的消融实验？还有就是什么时候开源SparseOcc❥(^_-)

#58 cheng-peng-git closed 4 months ago
2
unexpected key in source state_dict，missing keys in source state_dict

#57 zbaishancha closed 4 months ago
3
Using flip augmentation in the training, how to align the preds bbox(on the flip imgs) to the unflip gt bbox of original ann_file？

#56 Slonna closed 4 months ago
8
去噪任务的影响

#55 huichen98 closed 2 days ago
6
提Issue的各位哥哥姐姐们能不能先点个star啊

#54 afterthat97 opened 4 months ago
1
Hi, I want to know that what isthe gen_sweep_info.py used for? Does SparseBEV use nuscenes sweeps data to train and inference?

#53 Slonna closed 4 months ago
5
why there is a re-implementation of gradient checkpointing

#52 zen-d closed 2 days ago
3
Online Inference

#51 YoushaaMurhij closed 4 months ago
1
It seems that use the split val dataset for training besides the original split trainset under the nuscenes leaderboard submission config setting ?

#50 Slonna closed 4 months ago
2
Question about merging SparseBEV into the MMDection3D Framework.

#49 tualatinlz closed 2 days ago
1
Evaluation Error

#48 smritish21 closed 5 months ago
4
SparseOcc questions

#47 vehxianfish closed 5 months ago
5
Release model: vit_eva02_1600x640_trainval_future

#46 afterthat97 closed 5 months ago
0
Is there any reason for setting the maximum sampling point to 32?

#45 Qidian213 closed 6 months ago
2
Is there ablation study about the train_interval?

#44 tang-hy opened 6 months ago
2
Where in the code do you handle ego motion?

#43 Slonna closed 7 months ago
2
Question about absolute velo

#42 iMbaGong closed 7 months ago
2
simple_test_{online, offline} difference

#41 Lewis-Lu closed 7 months ago
2
img_curr_large = img[:, i] # [B, 6, C, H, W] IndexError: index 1 is out of bounds for dimension 1 with size 1

#40 matrix97317 closed 7 months ago
5
Cuda out of memory

#39 Orange-rightperson closed 2 days ago
2
关于三个模块

#38 GreatZeZuo closed 7 months ago
0
Can't find `sample_infos = pickle.load(open(os.path.join(args.data_root, 'nuscenes_infos_train.pkl'), 'rb'))`

#37 matrix97317 closed 7 months ago
2
预训练模型下载链接（国内网盘）

#36 CeoiZidung closed 7 months ago
6
Does sparseBEV support training with a single frame?

#35 Cc-Hy closed 7 months ago
2
V2-99 和Dual-Swin-T哪个主干网络好

#34 rubbish001 closed 7 months ago
2
Deployment questions

#33 BarathwajAnandan opened 7 months ago
1
关于预训练权重下载问题

#32 BarryGustin closed 2 days ago
4
About the generation of training data

#31 tang-hy closed 8 months ago
1
Question about more frames

#30 Debrove opened 8 months ago
5