Missmatched TensorRT version between nvcr.io/nvidia/deepstream-l4t:6.1-base and Jetpack 5.0.1 DP

Please provide complete information as applicable to your setup.

• Hardware Platform (Jetson / GPU)
• DeepStream Version
• JetPack Version (valid for Jetson only)
• TensorRT Version
• NVIDIA GPU Driver Version (valid for GPU only)
• Issue Type( questions, new requirements, bugs)
• How to reproduce the issue ? (This is for bugs. Including which sample app is using, the configuration files content, the command line used and other details for reproducing)
• Requirement details( This is for new requirement. Including the module name-for which plugin or for which sample application, the function description)

Hi, currently nvcr.io/nvidia/deepstream-l4t:6.1-base docker image has TensorRT 8.4.0.9 and Jetpack 5.0.1 DP has TensorRT 8.4.0.11.
Therefore, TensorRT engine built in Jetpack 5.0.1 environment can not be deployed with deepstream-l4t:6.1-base

You need to rebuild the engine on different device Jetson or dGPU.

Hi, in this page DeepStream-l4t | NVIDIA NGC
DeepStream 6.1 docker image is said that it support TensorRT 8.4.0.11, but actually TensorRT version inside the image is 8.4.0.9. The engine built natively in jetson host environment can not run inside docker environment due the TensorRT version issue, not device differences

How you check the version, I just downloaded docker image
nvcr.io/nvidia/deepstream-l4t:6.1-base
but tensorrt version is 8.4.0.1
root@b9ce584eb26d:/opt/nvidia/deepstream/deepstream-6.1# dpkg -l|grep nvinfer
ii libnvinfer-bin 8.4.0-1+cuda11.4 arm64 TensorRT binaries
ii libnvinfer-plugin8 8.4.0-1+cuda11.4 arm64 TensorRT plugin libraries
ii libnvinfer8 8.4.0-1+cuda11.4 arm64 TensorRT runtime libraries
ii python3-libnvinfer 8.4.0-1+cuda11.4 arm64 Python 3 bindings for TensorRT

Hello amycao,
I checked TensorRT version in the symbols of runtime nvinfer library
nm -D /usr/lib/aarch64-linux-gnu/libnvinfer.so | grep tensorrt_version

Can you build the engine inside docker?

Sure we can do that, but that will bring complexity to the current workflow. The docker deployment environment should be compatible with the development environment (native Jetson). I understand the Jetpack 5.0.1 is only developer preview version and wish to have a fix in the future releases.

Please check this FAQ, you can get the TRT 8.4.0.11 via APT
DeepStream SDK FAQ - Intelligent Video Analytics / DeepStream SDK - NVIDIA Developer Forums

There is no update from you for a period, assuming this is not an issue anymore.
Hence we are closing this topic. If need further support, please open a new one.
Thanks

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.