I am managing a high spec computer node with multiple GPUs, multiples cores and RAM. The idea is that users can share this resource by deploying Docker containers to run trainin