An Open-Source Large-Scale Reinforcement Learning Project for Search Agents
active 2025-08-07 → 2026-06-10 (UTC)
Activity over time
Daily event counts in the loaded window
Line chart, 308 days from 2025-08-07 to 2026-06-10. Pushes: 24 total, peak 3 in a day. Pull requests: 9 total, peak 2 in a day. Issues: 23 total, peak 4 in a day. Comments: 26 total, peak 4 in a day. Stars: 276 total, peak 43 in a day.
- Pushes
- Pull requests
- Issues
- Comments
- Stars
Stars, PRs, issues and forks are under-captured in the later part of this window. GH Archive progressively stopped capturing non-push events during 2026 — −95% or worse by the end of the window. Every series here except Pushes fades for that reason, so a decline above reflects the archive, not this repository. Pushes stay reliable throughout, so read them, and the contributor counts derived from them, as the real signal. Data health has the measurements.
Top contributors
Pushes, PRs, issues, reviews and comments — stars and forks excluded, so this is contribution rather than popularity
| Contributor | Contributions | Pushes | PRs | Comments |
|---|---|---|---|---|
| samjia2000 | 45 | 21 | 5 | 16 |
| GurrenLagann97 | 8 | 3 | 4 | 1 |
| daybreaksly | 3 | 0 | 0 | 1 |
| wenjunli-0 | 3 | 0 | 0 | 2 |
| Huziyou | 2 | 0 | 0 | 1 |
| cjz007 | 2 | 0 | 0 | 0 |
| Wenze7 | 2 | 0 | 0 | 1 |
| ligang-cs | 2 | 0 | 0 | 1 |
| bestpredicts | 2 | 0 | 0 | 0 |
| WujiangXu | 2 | 0 | 0 | 0 |
| Necolizer | 2 | 0 | 0 | 1 |
| unknowed-ER | 2 | 0 | 0 | 1 |
| G-JWLee | 1 | 0 | 0 | 0 |
| csung7 | 1 | 0 | 0 | 0 |
| MuL1ian | 1 | 0 | 0 | 0 |
| ZiyiTsang | 1 | 0 | 0 | 0 |
| jianshuod | 1 | 0 | 0 | 0 |
| HuichiZhou | 1 | 0 | 0 | 1 |
| NielsRogge | 1 | 0 | 0 | 0 |
Recent activity
Latest issues, pull requests and releases
- Issue comment#29wenjunli-02025-12-09 03:16Unstable Training
- Issue#31MuL1ian2025-11-01 09:51Question about LRM model training
- Issue comment#30wenjunli-02025-09-26 02:48Question regarding training steps
- Issue comment#30samjia20002025-09-26 01:26Question regarding training steps
- Issue#29wenjunli-02025-09-22 09:10Unstable Training
- Issue comment#26unknowed-ER2025-09-12 06:58Search_only flag in ASearcherWorkflow causing incorrect prompt template usage across local and web modes
- Issue comment#26samjia20002025-09-12 05:34Search_only flag in ASearcherWorkflow causing incorrect prompt template usage across local and web modes
- Issue#27NielsRogge2025-09-11 12:37Release ASearcher-v2 / QwQ-32B Agent on Hugging Face
- Issue#26unknowed-ER2025-09-11 03:37Search_only flag in ASearcherWorkflow causing incorrect prompt template usage across local and web modes
- Pull request#24samjia20002025-09-08 07:37
- Pull request#24GurrenLagann972025-09-08 02:39
- Issue#23csung72025-09-06 05:23Support for LoRA + Example YAML config updates
- Issue comment#22ligang-cs2025-09-05 04:22Unable to reproduce results of ASearcher-Web-QwQ on GAIA.
- Issue comment#21Wenze72025-09-04 16:49训练hang住
- Issue comment#22samjia20002025-09-04 07:28Unable to reproduce results of ASearcher-Web-QwQ on GAIA.
- Issue#22ligang-cs2025-09-03 13:29Unable to reproduce results of ASearcher-Web-QwQ on GAIA.
- Issue comment#21samjia20002025-09-03 05:58训练hang住
- Issue#21Wenze72025-09-02 08:26训练hang住
- Issue#20daybreaksly2025-08-31 02:27Question: Can training be resumed, and what else is required besides recover_start_step?
- Issue#20daybreaksly2025-08-30 14:43Question: Can training be resumed, and what else is required besides recover_start_step?
- Issue#7samjia20002025-08-29 07:32Evaluation framework
- Issue comment#18Huziyou2025-08-29 06:10Question about Multi-Turn Training with External Tools (e.g., Search R1)
- Issue comment#19samjia20002025-08-29 03:50Differences Between Index Files from index_builder.py and Search-R1
- Issue#12samjia20002025-08-29 03:46Unable to reproduce results for ASearcher-Local-7B evaluation
- Issue#19Necolizer2025-08-29 02:46Differences Between Index Files from index_builder.py and Search-R1
Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 276 stars here means stars gained during the window, not the repo's star count.