[cluster][doc] A tutorial for ML training with GPU on Kubernetes

[kevin85421]

trafficstars

Why are these changes needed?

This PR provides a guide for ML training with GPU on Kubernetes. The guide shows users how to build a Kubernetes cluster, create a ray cluster, and submit a PyTorch image training job. We will run Ray's PyTorch image training benchmark with a 1 gigabyte training set.

Related issue number

Checks

• [x] I've signed off every commit(by using the -s flag, i.e., git commit -s) in this PR.
• [x] I've run scripts/format.sh to lint the changes in this PR.
• [x] I've included any doc changes needed for https://docs.ray.io/en/master/.
• [x] I've made sure the tests are passing. Note that there might be a few flaky tests, see the recent failures at https://flakey-tests.ray.io/
• Testing Strategy
  • [ ] Unit tests
  • [ ] Release tests
  • [ ] This PR is not tested :(

[kevin85421]

cc @DmitriGekhtman

[DmitriGekhtman]

@ray-project/ray-docs could someone review? (Not sure if the above @ mention notifies the team members, but let's try.)

[kevin85421]

I have some small suggestions. This is a good first cut.

To make this great going forward, we would ideally have: (a) instructions for AWS too and (b) have this tested with Ray Release tests. Otherwise there is a high chance it will break over time.

@pcmoritz Thank you for the recommendations! I have opened two issues to track the progress of (1) AWS instructions (2) Ray Release Tests.