How to upgrade from CuDNN 7.0 to greater version (i.e. 7.1.5) on JetPack 3.2 w/o reflashing to JetPack 3.3

Hi,

Wondering if there’s any documentation on how to upgrade JetPack’s 3.2 CuDNN 7.0 to a higher version. I’m faced with the following error below and assume one remedy is to update CuDNN to a higher version, that my code/model was originally compiled on. I downloaded “cuDNN v7.3.1 Runtime Library for Ubuntu16.04 (Deb)” from https://developer.nvidia.com/rdp/cudnn-download, but later read on a discussion forum that TensorRT 3.0 GA doesn’t support any CuDNN versions other than the packaged 7.0 in JetPack 3.2. I spent a long time setting up a variety of packages on the Jetson, hesitant to reflash to 3.3 and thereby wipe out the total image. Welcome thoughts - thanks for your feedback.

2018-10-21 17:05:51.219853: E tensorflow/stream_executor/cuda/cuda_dnn.cc:324] Loaded runtime CuDNN library: 7.0.5 but source was compiled with: 7.1.5. CuDNN library major and minor version needs to match or have higher minor version in case of CuDNN 7.0 or later version. If using a binary install, upgrade your CuDNN library. If building from sources, make sure the library loaded at runtime is compatible with the version specified during compile configuration.
Segmentation fault (core dumped)

Hi,

Please reflash your device with JetPack3.3 to install cuDNN 7.1.

There are some dependencies on the CUDA driver.
cuDNN won’t work if being used cross JetPack version.

More, please noticed that the Jetson(ARM) package is only available in JetPack.
The package on our website is for the x86-based environment.

Thanks.