NVIDIA-AI-IOT/Lidar_AI_Solution

Issues

• libspconv.so FP32 inference support?

  #304 opened 14 days ago by HanyuWu
  0

• bevfusion cuda12 libspconv support？

  #303 opened 14 days ago by xiongda777
  0

• When kernel_size = (1,1,1) for SparseConv3d, the onnx inference result is different from the torch

  #302 opened 15 days ago by Icecream-blue-sky
  0

• use myself camera ，lidar calibration parameters and data, use public models，can use cuda-bevfusion?

  #301 opened 16 days ago by sen9910
  0

• The onnx output of bevfusion lidar backbone is different from the torch output

  #300 opened 17 days ago by Icecream-blue-sky
  1

• 预测我自己训练的模型

  #279 opened 3 months ago by evil-master
  2

• how debug into cuda kernel in vscode?

  #299 opened 21 days ago by zranguai
  0

• CUDA Runtime error cudaMemsetAsync(output_feature_, 0x00, volumn_output_ * sizeof(half), _stream) # an illegal memory access was encountered, code = cudaErrorIllegalAddress [ 700 ] in file CUDA-BEVFusion/src/bevfusion/camera-bevpool.cu:116

  #295 opened 22 days ago by Icecream-blue-sky
  3

• How to get 80 in geometry.geometry_dim = nvtype::Int3(360, 360, 80)?

  #298 opened 22 days ago by Jayden9912
  0

• the sparse convolution to onnx inference output does not match the torch output

  #297 opened a month ago by Xiao-Hu-Z
  1

• missing spconv in buidling cuda-bevfusion

  #296 opened a month ago by Jayden9912
  1

• Why conv.weight.permute(4, 0, 1, 2, 3) in fuse_bn function of CUDA-BEVFusion/qat/lean/funcs.py?

  #294 opened a month ago by Icecream-blue-sky
  2

• libspconv source code

  #281 opened 4 months ago by jongsik-moon
  4

• Why does Building CUDA_BEVFusion asks for libcublas.so.12, and not 11?

  #293 opened 2 months ago by Tony-syr
  1

• I have one former camera and lidar, which model should I choose

  #292 opened 2 months ago by Zhao-Qihao
  0

• What version of protobuf has been used to build CUDA_BEVfusion ?

  #291 opened 3 months ago by Tony-syr
  2

• AttributeError: 'torch._C.TensorType' object has no attribute 'with_sizes'

  #290 opened 3 months ago by Gj-master
  0

• the provided PTX was compiled with an unsupported toolchain

  #289 opened 3 months ago by jiafuzeng
  0

• How to export Swint model to onnx?

  #288 opened 3 months ago by jorgemn9
  0

• CUDA-BEVFusion sparsity

  #287 opened 3 months ago by wjx10210
  1

• why not use TensorRT for LidarBackbone model in CUDA-BEVFusion?

  #285 opened 3 months ago by fsi-kawatawa
  1

• add node cause result wrong!

  #286 opened 3 months ago by NotOnlyAI
  1

• Spconv Engine Implementation

  #256 opened 6 months ago by barrydoooit
  5

• export spconv onnx error

  #269 opened 5 months ago by aaa1233211
  1

• CUDA-BEVFusion: difference in the result of onnx and pytorch model

  #283 opened 4 months ago by Jayden9912
  5

• which ORIN chip for fps numbers?

  #280 opened 3 months ago by kurquhar
  1

• Could libspconv provide the IOBinding feature like TensorRT?

  #264 opened 3 months ago by ThomasMouse
  1

• ImportError: Lidar_AI_Solution/CUDA-BEVFusion/bevfusion/mmdet3d/ops/bev_pool/bev_pool_ext.cpython-38-x86_64-linux-gnu.so: undefined symbol: _ZN3c104impl8GPUTrace13gpuTraceStateE

  #284 opened 3 months ago by anteaglew2hcs
  0

• export ONNX error

  #272 opened 5 months ago by slayerlpj
  1

• export-transfuser erro

  #278 opened 4 months ago by evil-master
  1

• Is there some bug in TensorRT?

  #271 opened 5 months ago by bujianyiwang
  1

• How to run the nuscenes dataset with CUDA-BEVFusion?

  #274 opened 5 months ago by onlywl9598
  1

• Sensor selection

  #277 opened 4 months ago by suhansu
  1

• Support for H100 and sm_90

  #265 opened 5 months ago by maueki
  2

• spconv onnx

  #282 opened 4 months ago by GEN418
  0

• How to export the SwinT model from BEVFusion?

  #275 opened 5 months ago by slayerlpj
  2

• Failed to build model model/resnet50int8/build/head.bbox.layernormplugin.plan.

  #261 opened 6 months ago by Happynewyear000
  9

• When running the nuscenes validation set on Jeston Orin 32g, it keeps killing itself. What could be the reason?

  #276 opened 4 months ago by Xiao-Hu-Z
  1

• How to deploy BEVSegmentationHead

  #273 opened 5 months ago by whbdmu
  0

• Is there a version of Voxelization for float type point cloud

  #270 opened 5 months ago by nuaawh
  0

• Centerpoint Int8 quantization and deployment

  #268 opened 5 months ago by zxxzxy
  0

• an illegal memory access was encountered

  #267 opened 5 months ago by erzhu222
  3

• The outputs is all NAN

  #266 opened 5 months ago by chenzb0712
  0

• Does the point cloud range need to be symmetrical from zero?

  #262 opened 6 months ago by san9569
  1

• The number of voxels in Centerpoint and CUDA-CenterPoint did not match.

  #263 opened 6 months ago by mx2013713828
  0

• Only lidar or camera

  #260 opened 6 months ago by MarioUncle-Wang
  0

• I train centerpoint with kitti, so how can i get rpn_centerhead_sim.onnx and centerpoint.scn.onnx??

  #259 opened 6 months ago by ruishanyin
  0

• Jetson AGX Xavier with Jetpack 5.10 nvcc fatal : Unsupported gpu architecture 'compute_'

  #258 opened 6 months ago by HXB-1997
  1

• Dockerfile

  #257 opened 6 months ago by Ankitzanzmera
  0

• which model should I use?

  #255 opened 6 months ago by mx2013713828
  5