my fork of VAR
active 2024-04-04 → 2026-06-24 (UTC)
Activity over time
Daily event counts in the loaded window
Line chart, 812 days from 2024-04-04 to 2026-06-24. Pushes: 56 total, peak 13 in a day. Pull requests: 17 total, peak 3 in a day. Issues: 258 total, peak 8 in a day. Comments: 401 total, peak 15 in a day. Stars: 8,057 total, peak 665 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 |
|---|---|---|---|---|
| keyu-tian | 138 | 37 | 2 | 74 |
| iFighting | 111 | 14 | 2 | 56 |
| enjoyyi00 | 34 | 5 | 0 | 11 |
| Kumbong | 22 | 0 | 0 | 21 |
| YilanWang | 13 | 0 | 0 | 7 |
| luohao123 | 10 | 0 | 0 | 9 |
| AlexzQQQ | 9 | 0 | 0 | 7 |
| lxa9867 | 6 | 0 | 0 | 3 |
| moeinheidari7829 | 6 | 0 | 0 | 2 |
| Aibecool174 | 6 | 0 | 0 | 1 |
| daixiangzi | 6 | 0 | 0 | 3 |
| AlbertLin0 | 5 | 0 | 0 | 3 |
| duyuxuan1486 | 5 | 0 | 0 | 5 |
| krennic999 | 5 | 0 | 0 | 5 |
| ckczzj | 5 | 0 | 0 | 3 |
| daiyixiang666 | 5 | 0 | 0 | 5 |
| stephenhky | 4 | 0 | 0 | 2 |
| kl2004 | 4 | 0 | 0 | 3 |
| Junda24 | 4 | 0 | 0 | 2 |
| tanshuai0219 | 4 | 0 | 0 | 2 |
Recent activity
Latest issues, pull requests and releases
- Issue#179enjoyyi002026-02-07 04:17📢 Introducing SSG (Scaled Spatial Guidance) for Multi-Scale VAR Generation via Information-Theoretic Analysis
- Issue#179Youngwoo-git2026-02-06 05:40📢 Introducing SSG (Scaled Spatial Guidance) for Multi-Scale VAR Generation via Information-Theoretic Analysis
- Issue#178GengzeZhou2025-12-24 03:56🌟 Introducing SAR (Self-Autoregressive Refinement): stable self-rollout that unlocks robust post-training for Visual Autoregressive image generation.
- Issue#177wangtong6272025-11-27 09:26🌈 Introducing DiverseVAR, a training-free framework that unleashes the inherent generative diversity of Visual Autoregressive models while preserving fidelity and text–image alignment.
- Issue comment#45Jiawei8042025-10-05 07:46FID misalignment
- Issue comment#132TheQY712025-09-23 14:02Some doubts about quantization loss
- Issue#175LWang10162025-09-17 05:44can you share the multi-scale VQ-VAE training and evaluation code?
- Issue comment#143MingsYang2025-09-09 01:36VAR training
- Issue comment#143MingsYang2025-09-09 00:41VAR training
- Issue comment#59yichen-xie-ai2025-09-02 18:43There was no increase in speed after installing flash-attn and xformer.
- Issue comment#59Kumbong2025-08-31 09:54There was no increase in speed after installing flash-attn and xformer.
- Issue comment#59yichen-xie-ai2025-08-31 07:36There was no increase in speed after installing flash-attn and xformer.
- Issue#174Yikai-Wang2025-08-20 02:38Introducing "Next Visual Granularity Generation"
- Pull request#172ZeroRF2025-08-12 02:44
- Issue comment#171onepeachbiubiubiu2025-08-01 07:23how to modify VAR for image translation task ?
- Issue#171keruoya2025-07-23 10:43how to modify VAR for image translation task ?
- Issue comment#170ShinHyun-soo2025-07-20 14:14Any plans to release the VQVAE training code?
- Issue comment#28ShinHyun-soo2025-07-20 13:45Training code for VAE
- Issue comment#71Kumbong2025-07-08 21:15Computation and Memory Consumption When Training Models
- Pull request#168msoolee2025-06-26 07:00
- Issue comment#167JiayongO-O2025-06-17 06:16请教高手：我使用作者的vae、vard16_pth但是跑的结果accm只有acc_mean ≈5%，而作者论文中的是acc_mean ≈ 82.6%
- Issue#167zys4452025-06-13 08:22请教高手：我使用作者的vae、vard16_pth但是跑的结果accm只有acc_mean ≈5%，而作者论文中的是acc_mean ≈ 82.6%
- Issue comment#136BingoHS2025-05-28 07:42加载权重问题
- Issue#166StargazerX02025-05-27 02:54🚀 Introducing an efficient KV cache compression framework tailerd for next-scale prediction.
- Issue comment#126vanishingkazen2025-05-26 12:41Transformer每个尺度的输入为什么是整个f_hat而不是前一个尺度的特征
Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 8,057 stars here means stars gained during the window, not the repo's star count.