isaac_ros_dnn_inference icon indicating copy to clipboard operation
isaac_ros_dnn_inference copied to clipboard

Hardware-accelerated DNN model inference ROS 2 packages using NVIDIA Triton/TensorRT for both Jetson and x86_64 with CUDA-capable GPU

Results 12 isaac_ros_dnn_inference issues
Sort by recently updated
recently updated
newest added

Hello, I want to perform inference using multiple models in my design on Jetson devices. I came across this issue [here](https://github.com/NVIDIA-ISAAC-ROS/isaac_ros_dnn_inference/issues/31), but it only addresses the scenario of multiple inputs...

Hello everyone! I'm trying to use DLA core on jetson orin nx for running my model inside the tensor_rt node provided by Isaac Ros. This is the code i used...

Hi there. I was trying to deploy a pointpillars .onnx model using the dnn_inference package. The model takes three tensor lists as input. Normally, I should write an encoder that...

I was trying to run the triton node using container, however when it's `Loading extension: gxf/triton/libgxf_triton_ext.so`, it throws error that `libtritonserver.so: cannot open shared object file: No such file or...

Hi, On jetson Xavier NX, L4T: 35.4.1, Jetpack:5.1.2: Try to run command: cd /workspaces/isaac_ros-dev && \ colcon build --symlink-install && \ source install/setup.bash Got error messages: This may be promoted...

At the [Isaac ROS Documentation](https://nvidia-isaac-ros.github.io/repositories_and_packages/isaac_ros_dnn_inference/isaac_ros_dnn_image_encoder/index.html) the input_width is used twice. ROS Parameter | Type | Default | Description -- | -- | -- | -- input_image_width | uint16_t | 0...

documentation

Hello everyone, i'm trying to pass to the dnn_image_encoder node images coming from other topics, try remapping the /different topic other than /image, but the node is not subscribing to...

needs info

Hi, Is there a way to build Triton for a different version and replace it with the one in ```gxf```? I can see it uses JetPack 502, but I'm not...

enhancement

For models exported by tao toolkit v4.0 (yolov v4 for my case): It seems like `encoder_image_mean` and `encoder_image_stddev` needs to be set to 0 and 1/255 respectively to get any...

enhancement

I would like to add parameters to load custom plugin shared libraries like trtexec. Is that change acceptable? If so, I would like to create a PR.

enhancement