cuda versions and cli-command sequence

[shaked-sapir]

Hi there! I have two questions:

1. What Cuda versions are needed for running the code? on the server I use, there are DriverVersion=470.57.02 and CudaVersion=11.4 but when I run the command latplan mode hanoi 5 3 20 StateAE I get the following error:
   tensorflow.python.framework.errors_impl.InternalError: cudaGetDevice() failed. Status: CUDA driver version is insufficient for CUDA runtime version,
   so I wonder how can I get it to run. (also tried running ./train_propositions.sh first, but got the same error)
2. as for the commands mentioned above: what is the exact sequence to run, for example, hanoi with the arguments mentioned? am I doing it properly? In this question I mean both train and run already-trained models.

Thanks in advance

[aymeric75]

I use an nvidia A100 gpu, the advised docker image is nvcr.io/nvidia/tensorflow:23.03-tf1-py3