[RLlib] Remove unneeded args from offline learning examples

[ArturNiederfahrenhorst]

Why are these changes needed?

We largely got rid of replay buffers in the context of offline RL in RLlib, but the examples have not beed changed.

#26665

Checks

• [x] 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
  • [x] This PR is not tested :(

[ArturNiederfahrenhorst]

I am actually not a fan of learning_starts -> min_size. min_size seems to imply the size of the RB, not how many samples there are in the buffer.

This comment is directed at another PR that moves the learning_starts parameters but needed to be merged into this one because they interfere. The comment has been resolved in the other PR.

[sven1977]

Hey @gjoliver , could you approve this, so it can be merged, now that questions have been addressed?