🚀ReVisual-R1 is a 7B open-source multimodal language model that follows a three-stage curriculum—cold-start pre-training, multimodal reinforcement learning, and text-only reinforcement learning—to achieve faithful, concise, and self-reflective state-of-the-art performance in visual and textual reasoning.
active 2025-05-16 → 2026-04-20 (UTC)
Activity over time
Daily event counts in the loaded window
Line chart, 340 days from 2025-05-16 to 2026-04-20. Pushes: 9 total, peak 3 in a day. Pull requests: 0 total, peak 0 in a day. Issues: 12 total, peak 1 in a day. Comments: 14 total, peak 2 in a day. Stars: 103 total, peak 15 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 |
|---|---|---|---|---|
| CSfufu | 20 | 8 | 0 | 8 |
| wdp-007 | 3 | 0 | 0 | 2 |
| lrlbbzl | 2 | 0 | 0 | 1 |
| TU2021 | 2 | 0 | 0 | 2 |
| ziqipang | 2 | 0 | 0 | 1 |
| jun297 | 1 | 0 | 0 | 0 |
| XiaoYee | 1 | 1 | 0 | 0 |
| SxJyJay | 1 | 0 | 0 | 0 |
| yangdongchao | 1 | 0 | 0 | 0 |
| syyuan2021 | 1 | 0 | 0 | 0 |
Recent activity
Latest issues, pull requests and releases
- Issue#16SxJyJay2025-07-27 08:11Reproducing the reported results with open-sourced 47k cold-start dataset
- Issue#15guwen0072025-07-20 06:26Questions about hyperparameters of cold start
- Issue#13syyuan20212025-07-09 14:31RL training receipe
- Issue comment#10CSfufu2025-07-08 22:09Endless thinking
- Issue#12CSfufu2025-07-08 22:08Can't access the Demo
- Issue comment#12CSfufu2025-07-08 22:08Can't access the Demo
- Issue#7CSfufu2025-06-29 04:30Request to Open-Source GRAMMAR Dataset
- Issue#11yangdongchao2025-06-27 02:56Question about the MMRL.
- Issue#8CSfufu2025-06-25 04:36Low overall GPU Util rate in GRPO training
- Issue#9jun2972025-06-23 06:30Question about reproducing results
- Issue comment#7TU20212025-06-23 02:16Request to Open-Source GRAMMAR Dataset
- Issue comment#7TU20212025-06-23 02:16Request to Open-Source GRAMMAR Dataset
- Issue comment#7CSfufu2025-06-17 15:22Request to Open-Source GRAMMAR Dataset
- Issue comment#6CSfufu2025-06-17 15:22More detailed environment?
- Issue comment#5CSfufu2025-06-09 07:34Inference tends to not stop easily
- Issue#5CSfufu2025-06-09 07:34Inference tends to not stop easily
- Issue comment#5wdp-0072025-06-08 13:40Inference tends to not stop easily
- Issue comment#5CSfufu2025-06-08 07:16Inference tends to not stop easily
- Issue#5wdp-0072025-06-08 03:38Inference tends to not stop easily
- Issue comment#4wdp-0072025-06-07 15:58RuntimeError: CUDA driver error: invalid argument
- Issue#3lrlbbzl2025-06-07 03:50Inquiry about the details of Section 3.1
- Issue comment#3lrlbbzl2025-06-07 03:50Inquiry about the details of Section 3.1
- Issue comment#1CSfufu2025-06-06 05:48request for citation :)
- Issue comment#2ziqipang2025-06-05 19:13GPU Resources
- Issue comment#2CSfufu2025-06-05 15:08GPU Resources
Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 103 stars here means stars gained during the window, not the repo's star count.