You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I converted the sparse convolution backbone of voxel rcnn to onnx by modifying export-scn.py and exptool.py, and the conversion was successful without any errors. But when loaded as an engine, the output for voxels input is always 0. May I ask if there are any possible issues.
ps: without changing other parts, only replace the loaded onnx file with the onnx of bevfusion lidar backbone, and the output is normal.
The text was updated successfully, but these errors were encountered:
I converted the sparse convolution backbone of voxel rcnn to onnx by modifying export-scn.py and exptool.py, and the conversion was successful without any errors. But when loaded as an engine, the output for voxels input is always 0. May I ask if there are any possible issues.
ps: without changing other parts, only replace the loaded onnx file with the onnx of bevfusion lidar backbone, and the output is normal.
The text was updated successfully, but these errors were encountered: