[ICLR25] High-performance Image Tokenizers for VAR and AR
active 2024-10-02 → 2026-04-06 (UTC)
Activity over time
Daily event counts in the loaded window
Line chart, 552 days from 2024-10-02 to 2026-04-06. Pushes: 81 total, peak 19 in a day. Pull requests: 1 total, peak 1 in a day. Issues: 42 total, peak 4 in a day. Comments: 67 total, peak 5 in a day. Stars: 271 total, peak 18 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 |
|---|---|---|---|---|
| lxa9867 | 69 | 35 | 0 | 27 |
| qiuk2 | 63 | 46 | 0 | 13 |
| HalvesChen | 9 | 0 | 0 | 5 |
| Pride-Huang | 8 | 0 | 1 | 5 |
| aopolin-lv | 6 | 0 | 0 | 5 |
| LiaoLW | 6 | 0 | 0 | 3 |
| Changlin-Lee | 4 | 0 | 0 | 2 |
| BOB-ZX | 3 | 0 | 0 | 2 |
| JWZhao-uestc | 3 | 0 | 0 | 1 |
| SmileShaun | 3 | 0 | 0 | 1 |
| krennic999 | 2 | 0 | 0 | 1 |
| Ceveloper | 2 | 0 | 0 | 1 |
| zhouxingguang | 2 | 0 | 0 | 1 |
| iamlockelightning | 1 | 0 | 0 | 0 |
| liang-fen | 1 | 0 | 0 | 0 |
| NielsRogge | 1 | 0 | 0 | 0 |
| qyh3000 | 1 | 0 | 0 | 0 |
| jiachunjin | 1 | 0 | 0 | 0 |
| hizening | 1 | 0 | 0 | 0 |
| veroveroxie | 1 | 0 | 0 | 0 |
Recent activity
Latest issues, pull requests and releases
- Issue comment#28Ceveloper2025-09-10 09:25Code (RobustTok) for perturbed FID computation (pFID)
- Issue#28Ceveloper2025-09-09 14:06Code (RobustTok) for perturbed FID computation (pFID)
- Issue#27LiaoLW2025-08-06 04:00Question about training VAR with pre-trained MSVR10P2
- Issue#26liang-fen2025-06-27 07:11Do you have a recomended ratio to adjust the latents from your trained encoder models?
- Issue comment#25BOB-ZX2025-06-13 07:49Regarding multi-scale training VAE
- Issue comment#25lxa98672025-06-13 07:46Regarding multi-scale training VAE
- Issue comment#25BOB-ZX2025-06-13 07:36Regarding multi-scale training VAE
- Issue comment#25lxa98672025-06-12 20:57Regarding multi-scale training VAE
- Issue#25BOB-ZX2025-06-10 10:57Regarding multi-scale training VAE
- Issue comment#24lxa98672025-05-13 20:14training problem
- Issue comment#24aopolin-lv2025-05-13 19:56training problem
- Issue comment#24aopolin-lv2025-05-13 19:49training problem
- Issue comment#24qiuk22025-05-13 19:44training problem
- Issue comment#24aopolin-lv2025-05-13 19:30training problem
- Issue#16qiuk22025-05-02 03:50VQModel issue in training VAR
- Issue#24aopolin-lv2025-04-30 10:31training problem
- Issue comment#24aopolin-lv2025-04-30 10:31training problem
- Issue comment#24lxa98672025-04-30 03:07training problem
- Issue comment#24aopolin-lv2025-04-30 02:57training problem
- Issue comment#16qiuk22025-04-25 04:59VQModel issue in training VAR
- Issue comment#16LiaoLW2025-04-24 06:39VQModel issue in training VAR
- Issue comment#16lxa98672025-04-24 04:25VQModel issue in training VAR
- Issue comment#16LiaoLW2025-04-24 02:47VQModel issue in training VAR
- Issue comment#23LiaoLW2025-04-19 07:11Why quantizer mask is filled by value=si instead of value=1?
- Issue#23LiaoLW2025-04-19 07:11Why quantizer mask is filled by value=si instead of value=1?
Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 271 stars here means stars gained during the window, not the repo's star count.