Distribution of CUDA and cuDNN #inception

Hello,

My company, which is in the inception program, distributes a vision processing software for clients to be installed on their computers. We currently advice users to install CUDA and cuDNN SDK manually if they are to utilize AI modules in our application. However, we have received a lot of input on the hassle they’ve been through.

Thus, we are considering shipping binaries for both SDK’s. We’ve went through each requirement in both licenses:

https://docs.nvidia.com/cuda/eula/index.html#distribution-requirements

and saw that we could meet the requirements and ship binaries. Are there anyone with the same experience? Is there anything we’ve missed?