LLM Dataset & Evaluation Platform
14.0k 2026-04-18
ConardLi/easy-dataset
An application for generating high-quality datasets for LLM fine-tuning, RAG, and evaluation, featuring intelligent document processing and a comprehensive evaluation system.
Core Features
Intelligent Document Processing & Text Splitting
Multiple Dataset Types (QA, Dialogue, Image QA)
Automated Model Evaluation & Human Blind Test
Custom Prompts & Task Management
Multiple Export Formats (Alpaca, ShareGPT, Multilingual-T)
Detailed Introduction
Easy Dataset is a specialized application designed to streamline the creation of high-quality datasets for Large Language Models. It offers robust document parsing, intelligent segmentation, and data cleaning capabilities to transform diverse domain-specific documents into structured datasets. The platform supports various dataset types, including QA and dialogue, and integrates a comprehensive model evaluation system with automated scoring and human blind testing, making it invaluable for LLM fine-tuning, RAG, and performance assessment.