I was looking at the tensorRT SW license agreement, and there is a clause that I would like to get some clarity on.
It says “you agree that you will not (iii) sell, transfer, assign, distribute, rent, loan, lease, sublicense or otherwise make available the Licensed Software…”
Does that mean I can not distribute the customized tensorRT container for my purchased Nvidia devices?
If I use kubernetes framework to deploy my application on the customized tensorRT container, would it be a problem?
Hi,
I’ll have to look into this one. Stay tuned.
Thank you for your help. I am still looking forward to the answer.
Hi,
The answer I just got back was that “distribute” refers directly to this section: LICENSE AGREEMENT FOR NVIDIA SOFTWARE DEVELOPMENT KITS :: NVIDIA Deep Learning TensorRT Documentation. Please review this section to make sure you’re following the rules outlined there.
Thank you for the response. But I am still confused. Let assume that I made a ML/DL application included container leveraging tensorRT container(nvcr.io/nvidia/tensorrt:20.01-py3) and want to distribute the container for purchased jetson devices via our own distribution systems and private docker registry. Do you think it would be a SW license issue? The section you mention above, I found “you may further authorize Customers to redistribute the libnvinfer, libnvinfer_plugin, and libnvparsers libraries” but the tensorRT container.
I am wondering about the same thing. Maybe even a bit less private than @hyogrin. I have created a deep learning based algorithm that makes use of TensorRT and I want to make the source code publicly available. Since I want to make it as easy as possible for users to run the code, I plan to build a Singularity image that contains TensorRT. Would I be allowed to distribute such an image?