Is it possible to use tensorrt on a docker image based on python on a jetson nano? I know that L4T TensorRT exists. But those images are very big, Gigabytes for an image where I just run a simply python app
Hi,
You will still need all the dependencies installed inside the container (ex. CUDA, cuDNN, TensorRT).
So please use the L4T TensorRT image.
Thanks.
Is there at least a way to reduce that image size?
Hi,
You can try to build your own image on the top of l4t-base.
Thanks.