{"full_name":"microsoft/agent-lightning","owner":"microsoft","name":"agent-lightning","shard":"1","url":"https://github.com/microsoft/agent-lightning","description":"The absolute trainer to light up AI agents.","homepage":"https://microsoft.github.io/agent-lightning/","topics":["agent","agentic-ai","llm","mlops","reinforcement-learning"],"tag_pages":["agent","agentic-ai","llm","mlops","reinforcement-learning"],"language":"Python","license":"MIT","avatar":"https://avatars.githubusercontent.com/u/6154722?v=4","created_at":"2025-06-18T07:28:45Z","pushed_at":"2026-11-12T03:27:22Z","metrics":{"stars":18958,"forks":1723,"watchers":79,"issues":203,"prs":492,"releases":11,"issues_open":116,"issues_closed":87,"prs_open":68,"prs_closed":98,"prs_merged":326,"discussions":0},"growth":{"day":{"stars":5,"forks":1,"base_day":"2026-11-15"},"week":{"stars":70,"forks":17,"base_day":"2026-11-04"},"month":{"stars":288,"forks":53,"base_day":"2026-10-17"}},"history":{"dates":["2026-09-26","2026-09-27","2026-09-28","2026-09-29","2026-09-30","2026-10-01","2026-10-02","2026-10-03","2026-10-04","2026-10-05","2026-10-06","2026-10-07","2026-10-08","2026-10-09","2026-10-10","2026-10-11","2026-10-12","2026-10-13","2026-10-14","2026-10-15","2026-10-16","2026-10-17","2026-10-18","2026-10-19","2026-10-20","2026-10-21","2026-10-22","2026-10-23","2026-10-24","2026-10-25","2026-10-26","2026-10-27","2026-10-28","2026-10-29","2026-10-30","2026-10-31","2026-11-01","2026-11-02","2026-11-03","2026-11-04","2026-11-10","2026-11-11","2026-11-12","2026-11-13","2026-11-14","2026-11-15","2026-11-16"],"stars":[18197,18287,18336,18377,18427,18441,18453,18465,18472,18480,18521,18532,18546,18550,18551,18564,18570,18576,18596,18624,18644,18670,18680,18698,18716,18727,18733,18755,18767,18782,18788,18793,18800,18817,18828,18835,18843,18849,18874,18888,18935,18943,18947,18948,18949,18953,18958],"forks":[1599,1612,1615,1621,1629,1630,1633,1635,1636,1638,1646,1648,1650,1651,1654,1656,1656,1656,1662,1665,1666,1670,1674,1683,1693,1723,1754,1821,1813,1691,1692,1691,1693,1694,1694,1697,1699,1700,1702,1706,1720,1720,1722,1722,1721,1722,1723],"watchers":[84,84,84,86,85,85,85,85,85,86,86,86,86,86,86,84,84,84,84,84,84,84,84,85,85,84,84,84,84,84,84,84,79,79,79,79,78,78,78,78,null,80,80,80,80,80,79],"issues_open":[null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,117,116,116,116,116,116],"prs_open":[null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,83,69,69,68,68,68]},"boards":{},"same_owner":[{"full_name":"microsoft/markitdown","name":"markitdown","shard":"1","boards":{"daily":151,"weekly":176,"monthly":103,"total":36}},{"full_name":"microsoft/vscode","name":"vscode","shard":"1","boards":{"daily_watchers":21,"total":37}},{"full_name":"microsoft/PowerToys","name":"PowerToys","shard":"1","boards":{"daily_watchers":111,"total":81}},{"full_name":"microsoft/generative-ai-for-beginners","name":"generative-ai-for-beginners","shard":"1","boards":{"daily_forks":40,"total":98}},{"full_name":"microsoft/TypeScript","name":"TypeScript","shard":"1","boards":{"daily_watchers":120,"total":120}},{"full_name":"microsoft/terminal","name":"terminal","shard":"1","boards":{"total":133}},{"full_name":"microsoft/playwright","name":"playwright","shard":"1","boards":{"total":143}},{"full_name":"microsoft/Web-Dev-For-Beginners","name":"Web-Dev-For-Beginners","shard":"1","boards":{"total":155}},{"full_name":"microsoft/ML-For-Beginners","name":"ML-For-Beginners","shard":"1","boards":{"total":173}},{"full_name":"microsoft/AI-For-Beginners","name":"AI-For-Beginners","shard":"1","boards":{}},{"full_name":"microsoft/AirSim","name":"AirSim","shard":"1","boards":{}},{"full_name":"microsoft/BitNet","name":"BitNet","shard":"1","boards":{}},{"full_name":"microsoft/Bringing-Old-Photos-Back-to-Life","name":"Bringing-Old-Photos-Back-to-Life","shard":"1","boards":{}},{"full_name":"microsoft/CNTK","name":"CNTK","shard":"1","boards":{}},{"full_name":"microsoft/Data-Science-For-Beginners","name":"Data-Science-For-Beginners","shard":"1","boards":{}},{"full_name":"microsoft/FluidFramework","name":"FluidFramework","shard":"1","boards":{}},{"full_name":"microsoft/IoT-For-Beginners","name":"IoT-For-Beginners","shard":"1","boards":{}},{"full_name":"microsoft/MMdnn","name":"MMdnn","shard":"1","boards":{}},{"full_name":"microsoft/Ontology-Playground","name":"Ontology-Playground","shard":"1","boards":{}},{"full_name":"microsoft/Procexp-Mac","name":"Procexp-Mac","shard":"1","boards":{"daily_rate":96,"weekly_rate":32,"monthly_rate":87}}],"issues":[{"number":116,"title":"Based on the rag project, there are several questions","comments":26,"created_at":"2025-09-30T04:18:58Z"},{"number":401,"title":"How to save val_data predictions during each evaluation of `trainer.fit`?","comments":15,"created_at":"2025-12-11T06:58:41Z"},{"number":89,"title":"Is is possible to support Google Agent Development Kit (ADK) in the future?","comments":15,"created_at":"2025-09-12T06:22:57Z"},{"number":425,"title":"APO's TraceToMessages adapter fails with multi-turn agent rollouts (KeyError: 'role')","comments":14,"created_at":"2025-12-17T13:45:18Z"},{"number":119,"title":"I think dont put compute_response_mask in trainer.py, otherwise response_mask may not contiguous, put it in deamon.py?","comments":13,"created_at":"2025-09-30T10:21:56Z"},{"number":438,"title":"[BUG] Potential Memory Leak Issue","comments":12,"created_at":"2025-12-29T04:31:29Z"},{"number":352,"title":"Error with vllm","comments":11,"created_at":"2025-12-03T03:10:45Z"},{"number":337,"title":"🚀 Refactoring Proposal: Adopt Ruff for Integrated Linting and Formatting","comments":11,"created_at":"2025-11-26T08:28:55Z"}],"discussions":[],"latest_release":{"name":"Agent Lightning v1.0.3","tag":"v1.0.3","published_at":"2026-11-03T08:47:17Z","body":"## Highlights\r\n\r\n- **Score Centering for low-precision RL:** Added Score Centering with top-K behavior-probability capture and a runnable Qwen3.5-9B math training example. (#651, #677).\r\n- **Lower-memory MoE routing replay:** Reduced peak host-memory usage when decoding routed experts by processing one rollout at a time, while preserving the resulting training tensors (#636).\r\n\r\n## Features\r\n\r\n- Added INT8 weight/activation rollouts with an INT4 Attention KV cache, weight synchronization, and model preparation and precision-verification tools for the math training example (#651).\r\n- Added opt-in trajectory prefix-mismatch diagnostics to help identify possible retokenization mismatches without changing training tensors. Additional classification is disabled by default (#646).\r\n\r\n## Bug Fixes\r\n\r\n- Preserve fractional rewards during advantage computation and prevent precision loss or overflow during rollout normalization (#640, #662).\r\n- Preserve source log probabilities during trajectory aggregation, avoiding input mutation and inconsistent results when converting a rollout more than once (#613).\r\n- Correctly decode local image file URIs, including paths containing spaces, Unicode, or percent characters, so valid multimodal training rows are retained (#638).\r\n- Load custom tool-parser plugins before starting VERL's embedded vLLM server (#665).\r\n- Reject unknown top-level fields in rollout PATCH requests with HTTP 422 instead of silently ignoring them (#611).\r\n- Handle overflowing log probabilities in triplet queries without failing the request, while preserving token IDs (#647).\r\n- Report failed rollout-deletion requests instead of silently treating them as successful cleanup (#656).\r\n- Mark rollouts as failed when Kubernetes Job templates cannot be rendered or parsed, rather than retrying indefinitely (#657).\r\n\r\n## Contributors\r\n\r\nHuman contributors, listed alphabetically by GitHub username:\r\n\r\n@feiiiiii5, @galleonli, @hzy46, @kiteretsu903, @ldemon2333, @lll-hhh, @lux-liang, @NaCr05, @ningyixi, @original4422.\r\n","is_prerelease":false},"readme":true,"readme_zh":false,"news":[],"stale":false,"snapshot_at":"2026-11-16T09:23:14Z","history_start":"2026-09-06"}