ReSearch: Learning to Reason with Search for LLMs via Reinforcement Learning
active 2025-03-03 → 2025-04-23 (UTC)
Activity over time
Daily event counts in the loaded window
Line chart, 52 days from 2025-03-03 to 2025-04-23. Pushes: 7 total, peak 2 in a day. Pull requests: 4 total, peak 3 in a day. Issues: 60 total, peak 7 in a day. Comments: 94 total, peak 14 in a day. Stars: 648 total, peak 126 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 |
|---|---|---|---|---|
| AnselCmy | 62 | 7 | 3 | 44 |
| qilong-zhang | 12 | 0 | 0 | 4 |
| zfj1998 | 10 | 0 | 0 | 7 |
| Danee-wawawa | 8 | 0 | 0 | 2 |
| diablounbounded | 7 | 0 | 0 | 4 |
| JohnConnor123 | 7 | 0 | 0 | 5 |
| frankjoey2048 | 7 | 0 | 0 | 7 |
| BeastyZ | 6 | 0 | 0 | 2 |
| ericg108 | 5 | 0 | 0 | 3 |
| XIRZC | 5 | 0 | 0 | 2 |
| ccp123456789 | 4 | 0 | 0 | 0 |
| LePanda026 | 3 | 0 | 0 | 3 |
| zihaolucky | 3 | 0 | 0 | 3 |
| vincehxb | 3 | 0 | 0 | 0 |
| xiariyuni | 3 | 0 | 0 | 0 |
| NielsRogge | 2 | 0 | 0 | 1 |
| shuxiusuxiu | 2 | 0 | 0 | 0 |
| histmeisah | 2 | 0 | 0 | 0 |
| pengjeck | 2 | 0 | 1 | 1 |
| tiannuo-yang | 2 | 0 | 0 | 2 |
Recent activity
Latest issues, pull requests and releases
- Pull request#44AnselCmy2025-04-23 19:33
- Pull request#44AnselCmy2025-04-23 19:32
- Issue#43kehuitt2025-04-23 13:46sglang server
- Pull request#39AnselCmy2025-04-23 06:55
- Issue#41XIRZC2025-04-21 09:53Request for Key Changes in ReSearch Since verl-v0.2.0.post2 to Support Migration to verl-v0.3.0.post1
- Issue comment#39pengjeck2025-04-21 09:44feat(retriever): add thread pool and lock for concurrent safety
- Issue comment#37AnselCmy2025-04-18 15:57Dear author, regarding the issue of training speed
- Issue comment#33AnselCmy2025-04-18 15:54abut multi-node train
- Issue comment#38AnselCmy2025-04-18 15:30Training with an ERROR, and speed issue
- Issue comment#31JohnConnor1232025-04-18 15:13[2025-04-16 18:02:10,426 E 4034584 4034584] core_worker.cc:228: Failed to register worker 01000000ffffffffffffffffffffffffffffffffffffffffffffffff to Raylet. IOError: [RayletClient] Unable to register worker with raylet. No such file or directory
- Issue#40diablounbounded2025-04-18 12:06About running llm_judge.py
- Pull request#39pengjeck2025-04-18 08:57
- Issue#38ChrisRBXiong2025-04-18 06:48Training with an ERROR, and speed issue
- Issue#37xiariyuni2025-04-18 03:07Dear author, regarding the issue of training speed
- Issue#34xiariyuni2025-04-18 02:53运行到一半光闪烁但是不输出任何东西
- Issue#26qilong-zhang2025-04-18 02:35Evaluation Error
- Issue#28ccp1234567892025-04-18 00:59Project Details
- Issue comment#31JohnConnor1232025-04-17 22:18[2025-04-16 18:02:10,426 E 4034584 4034584] core_worker.cc:228: Failed to register worker 01000000ffffffffffffffffffffffffffffffffffffffffffffffff to Raylet. IOError: [RayletClient] Unable to register worker with raylet. No such file or directory
- Issue comment#36AnselCmy2025-04-17 17:45推理开了TP后的问题
- Issue comment#36dawson-chen2025-04-17 17:41推理开了TP后的问题
- Issue#35qilong-zhang2025-04-17 08:23How to mask <result>
- Issue comment#35qilong-zhang2025-04-17 07:59How to mask <result>
- Issue#35qilong-zhang2025-04-17 07:01How to mask <result>
- Issue#34xiariyuni2025-04-17 05:43报错：huggingface_hub.errors.HFValidationError: Repo id must be in the form 'repo_name' or 'namespace/repo_name': '/root/.cache/modelscope/hub/models/Qwen/qwen2.5-3b-instruct'. Use `repo_type` argument if needed.
- Issue#33Danee-wawawa2025-04-17 03:57abut multi-node train
Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 648 stars here means stars gained during the window, not the repo's star count.