question-mark
Stuck on an issue?

Lightrun Answers was designed to reduce the constant googling that comes with debugging 3rd party libraries. It collects links to all the places you might be looking at while hunting down a tough bug.

And, if you’re still stuck at the end, we’re happy to hop on a call to see how we can help out.

Voxel R-CNN on nuScenes dataset

See original GitHub issue

@djiajunustc Hi Jiajun, thanks for your excellent work of Voxel R-CNN. I have tried your Voxel R-CNN model on the datasets of KITTI and Waymo for multi-category detection. And the model performed very well. However, when I tried on the nuScenes dataset, I met the issue like this: File "/pcdet/models/roi_heads/target_assigner/proposal_target_layer.py", line 148, in sample_rois_for_rcnn batch_gt_of_rois[index] = cur_gt[gt_assignment[sampled_inds]] RuntimeError: The expanded size of the tensor (8) must match the existing size (10) at non-singleton dimension 1. Target sizes: [128, 8]. Tensor sizes: [128, 10].

I have tried to analyze and debug the code and model, so far there is no answer. I cannot figure out why the ground-truth’s code shape became to 10. Can you provide me with any enlightenment?

Issue Analytics

  • State:closed
  • Created 2 years ago
  • Comments:6 (1 by maintainers)

github_iconTop GitHub Comments

2reactions
deepmengcommented, Oct 3, 2021

Hi @deepmeng , recently, I’d like to use NuScenes datasets, but I am not sure about the performance of Voxel R-CNN. How are your results of Voxel R-CNN on NuScenes?

These are my evaluation results of Voxel R-CNN on the nuScenes dataset. image

0reactions
czy-0326commented, Oct 25, 2021

@deepmeng Hello, I try voxelrcnn on nuscenes dataset, and I get the same result as yours.

Read more comments on GitHub >

github_iconTop Results From Across the Web

Towards High Performance Voxel-based 3D Object Detection
Bearing this view in mind, we devise a simple but effective voxel-based framework, named Voxel R-CNN. By taking full advantage of voxel features...
Read more >
Towards High Performance Voxel-based 3D Object Detection
Voxel R-CNN consists of a 3D backbone network, a 2D bird-eye-view (BEV) Region Proposal Network, and a detect head. A voxel RoI pooling...
Read more >
Voxel Field Fusion for 3D Object Detection - Jiaya Jia
In this work, three different back- bones are adopted to validate the proposed framework, i.e.,. PV-RCNN [27] and Voxel R-CNN [8] on KITTI...
Read more >
arXiv:2002.10187v1 [cs.CV] 24 Feb 2020
We evaluate 3DSSD on widely used. KITTI dataset and more challenging nuScenes dataset. Our method outperforms all state-of-the-art voxel-based ...
Read more >
3D Object Detection | Papers With Code
Datasets · KITTI · nuScenes · ScanNet · NYUv2 · SUN RGB-D · S3DIS · Waymo Open Dataset · Argoverse ...
Read more >

github_iconTop Related Medium Post

No results found

github_iconTop Related StackOverflow Question

No results found

github_iconTroubleshoot Live Code

Lightrun enables developers to add logs, metrics and snapshots to live code - no restarts or redeploys required.
Start Free

github_iconTop Related Reddit Thread

No results found

github_iconTop Related Hackernoon Post

No results found

github_iconTop Related Tweet

No results found

github_iconTop Related Dev.to Post

No results found

github_iconTop Related Hashnode Post

No results found