README.md

metadata

title: DeepSeek OCR
emoji: 🚀
colorFrom: blue
colorTo: purple
sdk: gradio
sdk_version: 6.1.0
app_file: app.py
pinned: false
hardware: gpu

Ollama-Deepseek-OCR

A Gradio app for OCR using DeepSeek-OCR model from Hugging Face Transformers.

Setup

The app runs on Hugging Face Spaces with GPU support. No local setup required.

For local development:

1. Install dependencies: pip install -r requirements.txt

2. Run: python app.py

Features

• Convert documents to markdown
• Extract raw text
• Locate specific content with bounding boxes
• Support for images and PDFs

Note

This uses the DeepSeek-OCR model from Hugging Face Transformers for vision tasks. Supports both GPU and CPU inference.