🏆 Notable contributions
Popular open-source repos they've shipped to (commits + PRs)
Score breakdown
🛠 Featured work
Their own popular and pinned repositories
高性能计算相关知识学习笔记，包含学习笔记和相关知识的代码demo，在持续完善中。 如果有帮助的话请Star一下，对作者帮助很大，谢谢！
Jupyter NotebookCUDA C 编程权威指南代码实现 包含了书上第二章到第八章的大部分代码实现和作者笔记，全由作者本人手动实现，难免有错误的地方，请大家谨慎参考，非常欢迎对错误的指正。 如果有帮助的话请Star一下，对作者帮助很大，谢谢！
CudaWhisper inference with TensorRT-LLM
C++由作者收集整理的深度学习与机器学习电子书资料，包含多个不同方向，内容详尽种类多样（仅用于学习交流 如有侵权请联系 会立刻删除）
A baseline frame for recent papers including 10 different single-model supervised hashing methods, which can draw a mAP curves and Time curves for each methods in diverse bit number.
MATLABThis is a Toolkit uses five different algorithms to remove the reflection on single image.The results of the removal (including PSNR/SSIM and the image after reflection removal) will be obtained by using a web-application constructed by flask.
Python🧬 Stack & domains
🧬 Most similar developers
Closest profile, nearby score
🔥 Full roast
🔥 Your top original repo with 474 stars is just HPC learning notes, 18 fork repos are pure bookmarks, 20 external PRs closed by yourself, your GitHub resume is flashier than your actual engineering foot
Eddie-Wang1120 — 81.20/100 · ELITE (Elite · Top-tier dev)
TL;DR: Heavy-hitting core contributor to major open-source frameworks with a 99-PR merged track record, but your personal repo shelf is mostly study notes and bookmarks that can't carry the weight of your external work.
| Dimension | Score | Notes |
|---|---|---|
| Account maturity | 9.1/10 | Registered 5.25 years, active across 6 contribution years with zero days since last public activity |
| Original project quality | 7.2/18 | 11 original repos with 896 total stars, but top-starred original repo is HPC-Learning-Notes (474 stars, quality score 0.19, Jupyter notebook study notes) — your best-performing original project is a study notes repo, not a usable engineering project |
| Contribution quality | 24.2/27 | 99 merged PRs out of 133 total, 4% maintainer rejection rate (only 4 unmerged maintainer-closed PRs), 20 author-closed external PRs, 0 author-closed own-repo PRs — low rejection rate but occasional self-pruning of external contributions |
| Ecosystem / maintenance impact | 20/20 | 107 PRs + 82 commits across 6 popular repos with a combined 210k+ stars (including PaddlePaddle/Paddle, llama.cpp, TensorRT-LLM), 68 of 69 verified impact samples are core engineering changes (only 1 doc-like) |
| Community influence | 6.3/8 | 145 followers, following only 10 accounts — solid but unremarkable follower count, and your low following rate suggests you're here to ship code, not network |
| Activity authenticity | 14.4/17 | 173 contributions in the last year, only 7 trivial PRs in recent samples, 0 recent doc-like PRs, no templated PR flooding or bot-like behavior detected |
Red flags
- Your top-starred original repo (HPC-Learning-Notes, 474 stars) is a Jupyter notebook study notes collection with a quality score of 0.19, meaning nearly half of your original project star count comes from learning materials rather than usable engineering work.
- 20 of your external PRs were closed by you before merge, indicating occasional churn in your contribution workflow, though maintainer rejection rate remains low at 4%.
- 18 of your 29 public repos are forks, meaning nearly two-thirds of your GitHub presence is curated content from other projects rather than original work.
Score calibration No extra adjustment. The base score already prices in the low quality of your top-starred original repo and the heavy concentration of your impact on PaddlePaddle/Paddle.
Verdict Normal, high-value open-source contributor with a strong core engineering track record on major frameworks — your personal repo shelf just can't keep up with the quality of the work you do for other people's projects.