README.md

metadata

title: DocuMind
emoji: 🧠
colorFrom: blue
colorTo: purple
sdk: gradio
app_file: app.py
pinned: true
license: mit
sdk_version: 6.18.0
tags:
  - track:backyard
  - achievement:offgrid

🧠 DocuMind - Smart Document Analyzer

AI-powered document analysis and OCR tool | AI驱动的文档分析与OCR工具

✨ Features | 功能特性

Feature	Description	技术
📄 Document Analysis	Parse PDF, Word, TXT files	PyPDF2, python-docx
🤖 AI Summary	Generate intelligent summaries	GPT-2 (1.5B)
✍️ AI Rewrite	Rewrite with custom prompts	GPT-2 (1.5B)
💬 Document Q&A	Answer questions about document	GPT-2 + TF-IDF
🖼️ Image Description	AI image captioning	BLIP (0.4B)
🔍 OCR Recognition	Extract text from images	EasyOCR
🌐 Bilingual	Chinese & English support	-

🎬 Demo Video | 演示视频

• YouTube: https://youtu.be/1iR65sq0HmA

📱 Social Media | 社交媒体

• Twitter (Video): https://x.com/bingeaigc/status/2064999685476257999
• Twitter (Project): https://x.com/bingeaigc/status/2064999682582196665

🔗 Links | 链接

• Live Demo: https://huggingface.co/spaces/build-small-hackathon/documind
• GitHub: https://github.com/Binge-666/documind

🚀 Quick Start | 快速开始

pip install -r requirements.txt
python app.py
Open http://localhost:7860 | 打开 http://localhost:7860
📋 Requirements | 要求
- ✅ AI Model ≤ 32B params (GPT-2: 1.5B + BLIP: 0.4B)
- ✅ Hugging Face ecosystem
- ✅ Deployed on HF Spaces
🛠️ Tech Stack | 技术栈
- UI: Gradio
- NLP: GPT-2 (Hugging Face Transformers)
- Q&A: TF-IDF + Cosine Similarity (scikit-learn)
- Image AI: BLIP (Hugging Face Transformers)
- OCR: EasyOCR
- Document: PyPDF2, python-docx
📄 License
MIT