Log compaction-aware rollout metrics

[Shi-Dong]

Summary

• log rollout/num_training_samples for the finalized samples in each rollout step
• log rollout/episode_raw_reward by averaging samples within each original rollout, then averaging rollouts equally

Why

Session compaction can turn one rollout into many training samples. The existing raw-reward metric gives those rollouts proportionally more weight, and W&B does not currently expose how many samples a step trains on. These metrics make both effects visible without changing the meaning of existing metric histories.

Testing

• python -m pytest --confcutdir=tests/fast/ray/rollout tests/fast/ray/rollout/test_metrics.py -q (23 passed)
• uvx black --check miles/ray/rollout/metrics.py tests/fast/ray/rollout/test_metrics.py
• python -m ruff check miles/ray/rollout/metrics.py tests/fast/ray/rollout/test_metrics.py
• git diff --check

[claude]

Claude Code Review

This repository is configured for manual code reviews. Comment @claude review for a one-time review, or @claude review always to subscribe this PR to a review on every future push.

_{Tip: disable this comment in your organization's Code Review settings.}