I have a python machine learning script for wich I need special hardware (a 8-GPU machine or Tensor Processing units). Therefore I run the code on a cloud server.
Using Containers as known in docker looks like a very convenient way to execute the deep learning pytorch code. But there will likely be errors. How do I communicate (console output) with the docker process and get the resulting file out of the container. Do I need to upload (by network) it to a file server ?
Another question is how the access from a docker process to the Hardware of the machine works.
I need a CUDA Interface for docker and a cloud service machine with the appropriate hardware for deep learning applications (tensorflow, pytorch, …)
Are there best practices to run deep learning code with containers (at cloud services) ?
Thanks for your help and advice.
Source: Docker Questions