HopeJW

icon indicating a website link http://blog.csdn.net/hope_v icon indicating an email [email protected]

icon company NV icon location Suzhou, China

Results 128 comments of HopeJW

libspconv.so build

@guhuajun Sorry, I pushed [here](https://github.com/NVIDIA-AI-IOT/Lidar_AI_Solution/tree/master/libraries/3DSparseConvolution/libspconv-1.1.1), not a branch.

How to specifiy a custom protobuf path instead of default path?

I have updated the libspconv.so in a new strategy. We no longer require protobuf v3.6.1 as we can configure protobuf manually.

pytorch_quantization/cuda_ext.cpython-38-x86_64-linux-gnu.so: undefined symbol: _ZN3c1015SmallVectorBaseIjE8grow_podEPvmm

Hi, wei, You can refer the official site of pytorch_quantization_library to check it. ------------------ Original ------------------ From: Wei ZHANG ***@***.***> Date: Thu,Oct 26,2023 6:49 PM To: NVIDIA-AI-IOT/Lidar_AI_Solution ***@***.***> Cc: Subscribed...

About camera-geometry.cu

Is this wrong?

About camera-geometry.cu

Please give me some time, thansk.

CMake Compile Error

Oh, it looks like you are compiling in an abnormal environment.

CMake Compile Error

Sorry, this is caused by the option -std. For CUDA 11.x, we should set the option to -std=c++14. For CUDA 12.x, the option should be set to -std=c++17.

BEVSegmentationHead

Sorry, we have no plan to support BEVSegmentationHead currently.

Inference time 25 FPS

You should check your device freq. For example: run jetson_clocks.

Inference time 25 FPS

You should make sure your device is running on MAXN mode. ``` sudo nvpmodel -q NV Power Mode: MAXN ``` Change to MAXN can use the command: `sudo nvpmodel -m...