natesimon / MonoNav

MonoNav: MAV Navigation via Monocular Depth Estimation and Reconstruction
https://natesimon.github.io/mononav/
MIT License
90 stars 5 forks source link

Run Error #3

Closed dasda-asd closed 7 months ago

dasda-asd commented 7 months ago

This is a very good job. I run the fuse_dept. py file and the following error occurs.

The follow : [Open3D Error] (class std::shared_ptr cdecl open3d::core::CreateDeviceHashBackend(int64,const class open3d::core::Dtype &,const class open3d::core::SizeVector &,const class std::vector<class open3d::core::Dtype,class std::allocator > &,const class std::vector<class open3d::core::SizeVector,class std::allocator > &,const class open3d::core::Device &,const enum open3d::core::HashBackendType &)) D:\a\Open3D\Open3D\cpp\open3d\core\hashmap\DeviceHashBackend.cpp:38: Unimplemented device

Traceback (most recent call last): File "C:\Users\23860\Desktop\深度估计与SLAM融合\参考论文\MonoNav\MonoNav-main\fuse_depth.py", line 51, in vbg = VoxelBlockGrid(depth_scale, depth_max, trunc_voxel_multiplier, o3d.core.Device(device)) File "C:\Users\23860\Desktop\深度估计与SLAM融合\参考论文\MonoNav\MonoNav-main\utils\utils.py", line 50, in init self.vbg = o3d.t.geometry.VoxelBlockGrid( RuntimeError: [Open3D Error] (class std::shared_ptr cdecl open3d::core::CreateDeviceHashBackend(int64,const class open3d::core::Dtype &,const class open3d::core::SizeVector &,const class std::vector<class open3d::core::Dtype,class std::allocator > &,const class std::vector<class open3d::core::SizeVector,class std::allocator > &,const class open3d::core::Device &,const enum open3d::core::HashBackendType &)) D:\a\Open3D\Open3D\cpp\open3d\core\hashmap\DeviceHashBackend.cpp:38: Unimplemented device

I would appreciate it if you could reply to me!

natesimon commented 7 months ago

Hi @dasda-asd , this looks like a GPU issue to me. Could you confirm your device setting in config.yml?

  1. You can try setting device: "CPU:0" and retrying fuse_depth.py. If the code runs, it is likely the issue is with the GPU.
  2. Can you also check whether your GPU is recognized by your computer? (E.g., try running nvidia-smi in your terminal.) This is a good way to check whether your GPU is properly recognized.

I hope this helps!

natesimon commented 7 months ago

Hi @dasda-asd, was your issue resolved?