[CVPR 2025] DEIM: DETR with Improved Matching for Fast Convergence
active 2025-07-25 → 2026-05-15 (UTC)
Activity over time
Daily event counts in the loaded window
Line chart, 297 days from 2025-07-23 to 2026-05-15. Pushes: 3 total, peak 1 in a day. Pull requests: 3 total, peak 2 in a day. Issues: 9 total, peak 1 in a day. Comments: 40 total, peak 9 in a day. Stars: 251 total, peak 9 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 |
|---|---|---|---|---|
| ShihuaHuang95 | 10 | 0 | 0 | 10 |
| XiSHEN0220 | 5 | 3 | 1 | 1 |
| EwertzJN | 4 | 0 | 0 | 4 |
| Zoxive | 3 | 0 | 0 | 1 |
| wuruoyu1997coder | 2 | 0 | 0 | 2 |
| laxiyu | 2 | 0 | 0 | 1 |
| SebastianJanampa | 2 | 0 | 0 | 2 |
| ShihaoHan19980609 | 1 | 0 | 0 | 1 |
| brunogr-lotuslabs | 1 | 0 | 0 | 1 |
| LeBron-Jian | 1 | 0 | 0 | 1 |
| Wulingtian | 1 | 0 | 0 | 0 |
| xianbeisukisu | 1 | 0 | 0 | 1 |
| Unicorn123455678 | 1 | 0 | 0 | 1 |
| Hongru0306 | 1 | 0 | 1 | 0 |
| demuxin | 1 | 0 | 0 | 0 |
| surpoloyang | 1 | 0 | 0 | 1 |
| JiachengDeng | 1 | 0 | 0 | 1 |
| 1010happy | 1 | 0 | 0 | 0 |
| mouwumou | 1 | 0 | 0 | 1 |
| Bobbelty | 1 | 0 | 0 | 0 |
Recent activity
Latest issues, pull requests and releases
- Issue#130mo-lx2025-12-11 14:33deim-dfine训练时是否会加载预训练权重
- Issue comment#72wuruoyu1997coder2025-11-24 05:10AssertionError assert (boxes1[:, 2:] >= boxes1[:, :2]).all()
- Issue comment#15wuruoyu1997coder2025-11-24 04:40“nan" problem in training custom dataset
- Issue comment#15ShihuaHuang952025-11-01 01:11“nan" problem in training custom dataset
- Issue comment#108ShihuaHuang952025-11-01 01:08resume断点续训，接着训练loss从31增加到46，epoch训练时长从3小时增加5小时
- Issue comment#86ShihuaHuang952025-11-01 00:57Suggestion for requirements
- Issue comment#114ShihuaHuang952025-11-01 00:44当我配置好自己的数据集后，无法进行训练
- Issue comment#118ShihuaHuang952025-11-01 00:42torch.topk(scores.flatten(1), self.num_top_queries, dim=-1)，这行表示什么意思
- Issue comment#119ShihuaHuang952025-11-01 00:36多卡验证时，卡数不同验证的map不同
- Issue comment#123ShihuaHuang952025-11-01 00:29images = torch.cat([x[0][None] for x in items], dim=0)
- Issue comment#125ShihuaHuang952025-11-01 00:27请问有什么办法输出各个类别的AP吗
- Issue comment#126ShihuaHuang952025-11-01 00:24对于论文中Figure 2关于不同匹配方法中正负样本数量的疑惑。
- Issue comment#99JiachengDeng2025-10-31 07:21匹配感知损失(MAL)
- Issue#126leileilei20002025-10-30 13:15对于论文中Figure 2关于不同匹配方法中正负样本数量的疑惑。
- Issue#124Wulingtian2025-09-30 03:12Could you provide models from various series that were trained on the Objects365 dataset? They would be very helpful for fine-tuning. Thank you!
- Issue comment#122ShihuaHuang952025-09-28 07:01DEIM-D-FINE better than DEIM-RT-DETRv2 ？
- Issue comment#110JACK-Chen-20192025-09-26 04:03导出onnx报错
- Issue comment#103XiSHEN02202025-09-26 01:51DEIMv2 is on the way
- Issue#118demuxin2025-09-24 08:06torch.topk(scores.flatten(1), self.num_top_queries, dim=-1)，这行表示什么意思
- Issue comment#117laxiyu2025-09-19 03:19为什么不能修改模型的batch_size?
- Issue#117laxiyu2025-09-19 03:06为什么不能修改模型的batch_size?
- Pull request#116XiSHEN02202025-09-19 02:52
- Pull request#116Hongru03062025-09-19 00:59
- Issue comment#114mouwumou2025-09-16 22:03当我配置好自己的数据集后，无法进行训练
- Issue comment#103brunogr-lotuslabs2025-09-11 15:34DEIMv2 is on the way
Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 251 stars here means stars gained during the window, not the repo's star count.