Code for CRATE (Coding RAte reduction TransformEr). 马毅教授在神经网络可解释性方面的成果
active 2023-08-16 → 2026-03-06 (UTC)
Activity over time
Daily event counts in the loaded window
Line chart, 934 days from 2023-08-16 to 2026-03-06. Pushes: 65 total, peak 13 in a day. Pull requests: 7 total, peak 3 in a day. Issues: 35 total, peak 4 in a day. Comments: 19 total, peak 4 in a day. Stars: 931 total, peak 67 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 |
|---|---|---|---|---|
| DruvPai | 38 | 24 | 3 | 5 |
| yaodongyu | 30 | 23 | 0 | 4 |
| sdbuch | 13 | 9 | 2 | 2 |
| LeslieTrue | 13 | 8 | 0 | 4 |
| idonashino | 3 | 0 | 0 | 0 |
| sanwei111 | 3 | 0 | 0 | 0 |
| ElrondL | 3 | 0 | 0 | 1 |
| YanGe1105 | 2 | 0 | 0 | 1 |
| Carlson-J | 2 | 0 | 0 | 0 |
| LukasMahieu | 2 | 0 | 0 | 0 |
| ngkel | 2 | 0 | 0 | 0 |
| FJGEODEV | 1 | 0 | 0 | 0 |
| RobinWu218 | 1 | 1 | 0 | 0 |
| BiEchi | 1 | 0 | 1 | 0 |
| kachayev | 1 | 0 | 1 | 0 |
| 01vanilla | 1 | 0 | 0 | 0 |
| yiichu03 | 1 | 0 | 0 | 0 |
| happYLLLL | 1 | 0 | 0 | 1 |
| zhrli | 1 | 0 | 0 | 0 |
| heeseokjung | 1 | 0 | 0 | 0 |
Recent activity
Latest issues, pull requests and releases
- Issue#31Carlson-J2026-02-25 20:17Out Projection Matrix Different than Subspace Matrix
- Issue#31Carlson-J2026-02-19 21:20Out Projection Matrix Different than Subspace Matrix
- Issue#30sampath-ch2025-04-21 17:36Pre trained CRATE weights
- Issue comment#29YanGe11052024-07-31 07:08为什么我做分割效果和论文说的不一样，还没有vit好
- Issue#29YanGe11052024-07-31 07:07为什么我做分割效果和论文说的不一样，还没有vit好
- Issue#28LeslieTrue2024-07-26 14:40computing rate reduction in CRATE
- Issue comment#28LeslieTrue2024-07-26 14:40computing rate reduction in CRATE
- Issue comment#25yaodongyu2024-07-22 07:15the final task-specific architecture is classification head as the paper informed，but why you show demo in segmentation task？
- Issue#25yaodongyu2024-07-22 07:15the final task-specific architecture is classification head as the paper informed，but why you show demo in segmentation task？
- Issue comment#26yaodongyu2024-07-22 07:14can i implement this idea into some open source llm？such as qwen
- Issue#26yaodongyu2024-07-22 07:14can i implement this idea into some open source llm？such as qwen
- Issue comment#27yaodongyu2024-07-22 07:12where is the inference code？
- Issue#27yaodongyu2024-07-22 07:12where is the inference code？
- Issue#28heeseokjung2024-07-16 11:01computing rate reduction in CRATE
- Issue#27sanwei1112024-07-11 02:27where is the inference code？
- Issue#26sanwei1112024-07-05 08:46can i implement this idea into some open source llm？such as qwen
- Issue#25sanwei1112024-06-28 09:29the final task-specific architecture is classification head as the paper informed，but why you show demo in segmentation task？
- Issue#24zhrli2024-05-27 07:16How should I extract the features from unclassified data?
- Issue comment#8emanuelevivoli2024-04-26 10:51more pretrained weights
- Issue comment#23sdbuch2024-03-17 20:37Website update/reorganization
- Issue comment#23DruvPai2024-03-16 21:45Website update/reorganization
- Issue comment#23sdbuch2024-03-16 21:33Website update/reorganization
- Pull request#23DruvPai2024-03-16 02:59
- Pull request#22sdbuch2024-03-15 21:24
- Pull request#22sdbuch2024-03-15 21:24
Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 931 stars here means stars gained during the window, not the repo's star count.