verl

[doc] fix: `actor_rollout_ref.critic` is not correct (#3778)

[docker] feat: update Dockerfile.rocm7 (#3781)

[recipe] feat: Add example for gpt-oss training using agent loop (#3774)

[misc] feat: bump version to 0.7.0.dev (#3772)

[recipe] feat: add open math reasoning (#3767)