The code for paper 'Learning from Videos for 3D World: Enhancing MLLMs with 3D Vision Geometry Priors'
Jupyter Notebook · active 2025-06-02 → 2026-05-16 (UTC)
Activity over time
Daily event counts in the loaded window
Line chart, 349 days from 2025-06-02 to 2026-05-16. Pushes: 20 total, peak 4 in a day. Pull requests: 0 total, peak 0 in a day. Issues: 56 total, peak 6 in a day. Comments: 75 total, peak 5 in a day. Stars: 112 total, peak 5 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 |
|---|---|---|---|---|
| zd11024 | 72 | 20 | 0 | 41 |
| flyyyyer | 8 | 0 | 0 | 6 |
| Germany321 | 8 | 0 | 0 | 2 |
| missTL | 7 | 0 | 0 | 0 |
| 0x4e8 | 6 | 0 | 0 | 6 |
| Zhongan-Wang | 5 | 0 | 0 | 3 |
| Fusang-Wang | 4 | 0 | 0 | 2 |
| soul10101 | 4 | 0 | 0 | 1 |
| wpd-mlb | 3 | 0 | 0 | 2 |
| gaoxiangjun | 2 | 0 | 0 | 1 |
| hxiaoj | 2 | 0 | 0 | 0 |
| SuhZhang | 2 | 0 | 0 | 1 |
| npurson | 2 | 0 | 0 | 1 |
| Dlonly904 | 2 | 0 | 0 | 2 |
| Chongyu-hub | 2 | 0 | 0 | 1 |
| NeutrinoLiu | 1 | 0 | 0 | 0 |
| ZestfulJX | 1 | 0 | 0 | 0 |
| Terry-Xu-666 | 1 | 0 | 0 | 0 |
| ChengJade | 1 | 0 | 0 | 0 |
| PaulLiuYZ | 1 | 0 | 0 | 0 |
Recent activity
Latest issues, pull requests and releases
- Issue comment#68CSPlayer06892026-04-25 15:58训练日志的记录的loss和grad_norm与使用的GPU数量有关，这是Bug吗？
- Issue comment#65SuhZhang2026-04-24 03:37OOM when training on spatial reasoning with the default configuration
- Issue comment#68npurson2026-03-11 11:32训练日志的记录的loss和grad_norm与使用的GPU数量有关，这是Bug吗？
- Issue#65SuhZhang2026-02-17 03:40OOM when training on spatial reasoning with the default configuration
- Issue comment#64zd110242026-02-09 10:01是否可以公布spar_234k的媒体数据
- Issue comment#640x4e82026-02-09 09:29是否可以公布spar_234k的媒体数据
- Issue comment#640x4e82026-02-09 08:09是否可以公布spar_234k的媒体数据
- Issue comment#640x4e82026-02-09 07:53是否可以公布spar_234k的媒体数据
- Issue comment#640x4e82026-02-09 05:29是否可以公布spar_234k的媒体数据
- Issue#63ShiqiangLang2026-02-05 13:39Data construction problem
- Issue#61ChengJade2025-12-30 14:58Performance reproduce
- Issue#59npurson2025-12-30 10:54High grad_norm values during training and performance gap with my implementation
- Issue comment#59Dlonly9042025-12-29 03:59High grad_norm values during training and performance gap with my implementation
- Issue#56zd110242025-12-20 04:35About training speed
- Issue comment#56zd110242025-12-20 04:35About training speed
- Issue comment#57zd110242025-12-20 04:34Performance differences on visual grounding
- Issue comment#56Chongyu-hub2025-12-19 02:35About training speed
- Issue comment#56zd110242025-12-19 02:20About training speed
- Issue#56Chongyu-hub2025-12-18 03:49About training speed
- Issue comment#52zd110242025-11-20 07:49Media Data
- Issue comment#520x4e82025-11-20 07:33Media Data
- Issue comment#520x4e82025-11-20 06:19Media Data
- Issue comment#49zd110242025-11-19 15:36About frame sampling strategies
- Issue comment#15Dlonly9042025-11-18 09:28No matching distribution found for pytorch3d
- Issue#53Germany3212025-11-15 04:29Training loss difference
Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 112 stars here means stars gained during the window, not the repo's star count.