腾讯混元出品，模型压缩，Model compression toolkit engineered for enhanced usability, comprehensiveness, and efficiency.
Python · active 2025-07-04 → 2026-07-29 (UTC)
Activity over time
Daily event counts in the loaded window
Line chart, 391 days from 2025-07-04 to 2026-07-29. Pushes: 153 total, peak 8 in a day. Pull requests: 145 total, peak 8 in a day. Issues: 44 total, peak 4 in a day. Comments: 109 total, peak 6 in a day. Stars: 269 total, peak 17 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 |
|---|---|---|---|---|
| yghstill | 216 | 55 | 30 | 42 |
| liusong1222 | 51 | 14 | 23 | 8 |
| irisliu10 | 45 | 14 | 20 | 6 |
| github-actions[bot] | 36 | 36 | 0 | 0 |
| ali-88123 | 26 | 8 | 10 | 4 |
| WOODchen7 | 25 | 7 | 12 | 1 |
| RuBing-Yang | 23 | 6 | 13 | 2 |
| StromNoNo | 14 | 0 | 13 | 0 |
| GGgary666 | 10 | 2 | 2 | 4 |
| linchuanxie | 7 | 1 | 5 | 0 |
| yisunlp | 4 | 2 | 1 | 0 |
| dawnranger | 4 | 0 | 3 | 1 |
| loloto | 4 | 0 | 0 | 2 |
| Oldpan | 3 | 0 | 0 | 1 |
| lishaoao | 3 | 0 | 1 | 1 |
| panghongwei17 | 3 | 0 | 0 | 3 |
| ggg-s | 3 | 0 | 0 | 2 |
| Copilot | 3 | 0 | 0 | 3 |
| github-advanced-security[bot] | 3 | 0 | 0 | 2 |
| dynamicheart | 3 | 0 | 2 | 1 |
Recent activity
Latest issues, pull requests and releases
- Issue comment#318sunnyxiaohu2026-06-08 07:48fix(gptq): fix Hessian computation, variable-length sequence support, and layer output type handling
- Pull request#324zhzihao2026-05-30 08:49
- Issue#323kdrew9912026-05-29 20:07Source Code for Android Demo App?
- Pull request#319sunnyxiaohu2026-05-28 03:19
- Pull request#313yisunlp2026-05-27 03:45
- Issue comment#306chfjf2026-05-17 10:39[Bug] HY-MT-1.5-1.8B 1.25bit Incorrect Outputs
- Issue#306chfjf2026-05-15 00:03[Bug] HY-MT-1.5-1.8B 1.25bit Incorrect Outputs
- Pull request#303eppaneamd2026-05-14 04:59
- Issue comment#300irisliu102026-05-11 05:34请问下当前是否支持VLM模型的QAT训练
- Pull request#289linchuanxie2026-04-23 06:18
- Issue comment#287liusong12222026-04-20 09:15What was the training dataset/data mix used to train AngelSlim/Qwen3-8B_eagle3 and other qwen eagle3 drafters?
- Issue#287nelaturuharsha2026-04-19 06:38What was the training dataset/data mix used to train AngelSlim/Qwen3-8B_eagle3 and other qwen eagle3 drafters?
- Pull request#282gavingavin992026-04-15 08:08
- Issue comment#283yghstill2026-04-12 14:22为什么qwen3 1.7b eagle3 中文的猜中率很低，甚至猜不中。在线求?
- Pull request#280dawnranger2026-04-01 05:14
- Issue comment#277WOODchen72026-03-31 03:15Questions about QAT Data Composition
- Pull request#266irisliu102026-03-20 03:38
- Issue comment#263dawnranger2026-03-16 12:50RuntimeError: The size of tensor a (3) must match the size of tensor b (12) at non-singleton dimension 0
- Pull request#260dawnranger2026-03-16 08:38
- Issue#246yghstill2026-03-12 03:29pyarrow.lib.ArrowInvalid: offset overflow while concatenating arrays
- Pull request#254dawnranger2026-03-10 04:52
- Issue comment#230TheUgliestOfWuli2026-03-06 03:48[Eagle-3 Training] Cloud I use this code for Qwen3-235B draft model training. I can`t find any config about target model`s parallel setting like tp\pp\ep ...
- Pull request#251liusong12222026-03-04 06:53
- Pull request#249irisliu102026-03-03 11:17
- Pull request#247irisliu102026-02-28 06:49
Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 269 stars here means stars gained during the window, not the repo's star count.