search

NVIDIA-AI-IOT / Lidar_AI_Solution

A project demonstrating Lidar related AI solutions, including three GPU accelerated Lidar/camera DL networks (PointPillars, CenterPoint, BEVFusion) and the related libs (cuPCL, 3D SparseConvolution, YUV2RGB, cuOSD,).
Other
1.25k stars 218 forks source link

Train and test on custom data. #177

Open krishnatoshniwal opened 12 months ago

krishnatoshniwal commented 12 months ago

I want to test the Centerpoint model on some point cloud data that I have. Here's an outline of the steps that I'm currently thinking is required.

  1. Train the model with the model definition given in https://github.com/tianweiy/CenterPoint/blob/master/configs/nusc/voxelnet/nusc_centerpoint_voxelnet_0075voxel_fix_bn_z.py
  2. Export the model to onnx using - export-scn.py and export_neck_head.py.

Is that right? Also, what about models trained using mmdet3d, can the sparse conv present in these models also be converted to onnx using export-scn ?

hopef commented 11 months ago

yep, mmdet3d also supported by export-scn.

ruishanyin commented 2 months ago

Did u transfer your custom data to nuscenes? How?