[Tune] Fix HEBO evaluated rewards for max mode & save/restore

[Yard1]

Why are these changes needed?

If HEBO is set to maximize, evaluated rewards need to be multiplied by -1. Apologies, didn't catch that before.

Furthermore, this PR also makes sure that all the necessary attributes are saved and restored.

Checks

• I've run scripts/format.sh to lint the changes in this PR.
• I've included any doc changes needed for https://docs.ray.io/en/master/.
• 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 :(

[Yard1]

CC @krfricke

[krfricke]

Can you run ./scripts/format.sh once to run the linter?

[Yard1]

@krfricke I've also pushed a commit that'll make sure that all the necessary attributes are saved and restored - that should fix that issue on Slack, at least for HEBO.