waterkokoro

waterkokoro / eval-dominator

Public

基于OpenCompass的可视化评测系统。Visual Evaluation System Based on OpenCompass

14
0
100% credibility
Found May 02, 2026 at 14 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Vue
AI Summary

Eval Dominator is a lightweight local web app that provides a browser interface for evaluating large language models using OpenCompass.

How It Works

1
📥 Discover Eval Dominator

You find this free tool that lets you test AI models easily right in your web browser, without needing expert skills.

2
⚙️ Set it up simply

Follow quick preparation steps on your computer to get everything ready in minutes.

3
🔑 Log in quickly

Use the built-in starter account to access your personal dashboard instantly.

4
🤖 Add your AI model

Paste details for your favorite AI like ChatGPT or others, with your private info kept safe and hidden.

5
📊 Pick a test collection

Choose from ready example sets for math problems, questions, or knowledge tests.

6
▶️ Run the test

Hit start and watch your model tackle the challenges in real time with live updates.

✅ Review your results

Celebrate seeing clear scores, progress charts, logs, and downloadable files to understand how well it did.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 14 to 14 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is eval-dominator?

Eval-dominator is a local-first web system built with Vue that turns OpenCompass into a browser-driven eval platform for LLMs. You get a clean UI to submit evaluations against OpenAI-compatible APIs, pick from auto-synced OpenCompass demo datasets like GSM8K or CMMLU, track task progress with live logs and metrics tables, and download artifacts—all orchestrated via a Go backend and Python gRPC core. It solves the hassle of CLI-only OpenCompass by adding presets for models, task search/filtering, and cancel-anytime control.

Why is it gaining traction?

It stands out by making OpenCompass dead simple for quick remote API evals—no more manual config generation or log grepping. Features like masked API keys, auto-tail live infer logs, and percent-bar metrics hit dev pain points directly, while staying lightweight and offline-capable. Bilingual English/Chinese support broadens appeal without bloat.

Who should use this?

ML researchers benchmarking LLMs on OpenCompass datasets via APIs like DashScope or vLLM. Solo devs prototyping eval pipelines locally, skipping script boilerplate. Teams needing a visual dominator for single-machine eval tasks before scaling.

Verdict

Grab it for personal OpenCompass evals—solid MVP with intuitive UX and easy setup via scripts. At 14 stars and 1.0% credibility, it's early but iterating fast; stick to local single-user until multi-tenant lands. (187 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.