search

dusty-nv / jetson-inference

Hello AI World guide to deploying deep-learning inference networks and deep vision primitives with TensorRT and NVIDIA Jetson.
https://developer.nvidia.com/embedded/twodaystoademo
MIT License
7.85k stars 2.98k forks source link

libnvinfer.so: undefined reference to `nvdla::IProfile::setUseSoftMaxOptz(bool)' #1657

Open ElenaSkrebneva opened 1 year ago

ElenaSkrebneva commented 1 year ago

Hei, I have another issue with the "make" comment. Before you've just helped me solve an issue with "libcudla" (https://github.com/dusty-nv/jetson-inference/issues/1653) and I tried to continue with the "make", but got another one: /usr/lib/gcc/aarch64-linux-gnu/7/../../../aarch64-linux-gnu/libnvinfer.so: undefined reference to nvdla::IProfile::setUseSoftMaxOptz(bool)' /usr/lib/gcc/aarch64-linux-gnu/7/../../../aarch64-linux-gnu/libnvinfer.so: undefined reference tonvdla::IPoolingLayer::setPoolingPaddingInclusionType(nvdla::PoolingPaddingInclusionType)' Screenshot 2023-05-29 132012 Could you please try to help me solve this one, too? I have Jetson Nano SUB 4G version and Jetpack is 5.1.1-b56

dusty-nv commented 1 year ago

I have Jetson Nano SUB 4G version and Jetpack is 5.1.1-b56

Hi @ElenaSkrebneva, - if you are using Jetson Nano 4GB, I'm not sure how it has JetPack 5.1.1, because Jetson Nano supports JetPack 4.x. So you may need to use SD card with JetPack 4.6.3 or use SDK Manager to flash it. The errors you are getting seem to indicate a mismatch between the TensorRT version and the underlying GPU drivers in JetPack-L4T.

ElenaSkrebneva commented 1 year ago

It has it, because I've installed it (originally there was no JetPack in this SD card), using official instructions on: https://developer.nvidia.com/embedded/jetpack and installation went through. I didn't know about this compatibility issue. I will try to reinstall JetPack then I guess to the older version.

dusty-nv commented 1 year ago

@ElenaSkrebneva I think because you have the 3rd-party SUB model of Jetson Nano, maybe it came with different SD card image? With a fresh SD card image for it, you should be able to do sudo apt-get update && sudo apt-get install nvidia-jetpack if it doesn't come with JetPack.