Release v0.1.7

What's Changed

• Dependencies: Bump NeMo AutoModel to cfe722da and vLLM to 0.28, adding support for Qwen3.8-Flash-Next and GLM-5.3-Flash — #98 by @hijkzzz
• CI: Bump SSO preflight to v1.8.10 (AUT-1501) — #81 by @svcnemo-autobot
• Evaluation: Count each episode only once when computing pass@k — #84 by @hijkzzz
• Environment: Ensure proper teardown when closing environments — #83 by @hijkzzz
• Rollout: Prevent epochs from hanging when the dataset size is not divisible by rollout.batch_size — #85 by @TailinZhou
• Rollout: Improve epoch-tail handling — #86 by @hijkzzz
• Docker: Fix docker_run.sh crashes and prevent HF_TOKEN exposure — #89 by @bogdansalyp
• Documentation: Fix Qwen3-4B RL script references in the quickstart guide — #93 by @Wrestle7

New Contributors

• @TailinZhou made their first contribution in #85
• @bogdansalyp made their first contribution in #89
• @Wrestle7 made their first contribution in #93

Full Changelog: v0.1.6...v0.1.7