Hi,
I’m working on a Jetson-TX2 device with tensorrt7.1 installed in it. Now I want to test something in Tensorrt6.
How should I install it?
I tried using different tags of l4t and deepstream-l4t. But for all of them, I’m able to get only different versions of CUDA but not tensorrt.
Thanks in advance.
Hi,
Please reflash your device with JetPack4.3 directly.
There are some dependencies between GPU driver and libraries.
This will limit you to use the OS and package from the same JetPack rather than install an individual package.
Thanks.
1 Like
Same as docker use case.
Since L4T docker mounts the CUDA library from host, it is required to use the same software within and outside the docker.
Please reflash the device with JetPack4.3 and try 4.0.2-19.12-base here:
Thanks.
1 Like
@AastaLLL. Since other colleagues are using it for their experiments, I won’t be able to reflash the device. But anyway thanks for your useful comment.