if the input size of the spconv onnx should be dynamic ?

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

1.21k stars 206 forks source link

if the input size of the spconv onnx should be dynamic ? #220

Open pegasus-zhang opened 7 months ago

pegasus-zhang commented 7 months ago

hi, in this project the onnx file generated from spconv has a static input size of 51, while the generated trt engine from the onnx file shoud have an input size of voxelNumber5 which is dynamic , is there any conversion for this in the "Inference function for sparse convolution" in libspconv ?

hopef commented 7 months ago

By default, libspconv is running on dynamic input size. The input size (5, 1) does not affect the pipeline. It's just a placeholder!