# 🚀 Qwen OCR Hugging Face Deployment Package

Your Qwen OCR app is ready for deployment to Hugging Face Spaces!

## 📁 Package Contents

- **app.py** - Main Gradio application (API-only version, optimized for HF Spaces)
- **requirements.txt** - Minimal dependencies for the app
- **README.md** - Hugging Face Space card with app description
- **deploy.sh** - Automated deployment script (updated with better auth handling)
- **auth_check.sh** - Authentication status checker
- **deploy_instructions.md** - Detailed deployment instructions
- **test_local.py** - Local testing script

## 🔑 Authentication Required!

### First, check your authentication:
```bash
cd /Users/sangmin/Developer/Qwen2.5-VL/huggingface-space
./auth_check.sh
```

### Get your Hugging Face token:
1. Go to https://huggingface.co/settings/tokens
2. Create a new token with **WRITE** access (not READ!)
3. Copy the token (starts with `hf_...`)

## 🎯 Quick Deploy

### Option 1: Automated Script with Token (Recommended)
```bash
cd /Users/sangmin/Developer/Qwen2.5-VL/huggingface-space
./deploy.sh YOUR_HF_USERNAME qwen-ocr-detection YOUR_HF_TOKEN
```

Example:
```bash
./deploy.sh Sangmin qwen-ocr-detection hf_xxxxxxxxxxxxxxxxxxxx
```

### Option 2: Manual Web Upload
1. Go to https://huggingface.co/new-space
2. Create a new Gradio Space
3. Upload `app.py`, `requirements.txt`, and `README.md`
4. Wait for automatic deployment

### Option 3: Manual Git Commands
```bash
cd /Users/sangmin/Developer/Qwen2.5-VL/huggingface-space

# First, login with your token
huggingface-cli login
# Or set token as environment variable
export HF_TOKEN=your_token_here

# Create Space and push
huggingface-cli repo create YOUR_USERNAME/qwen-ocr-detection --type space --space_sdk gradio
git init
git add .
git commit -m "Initial commit"
git remote add origin https://huggingface.co/spaces/YOUR_USERNAME/qwen-ocr-detection
git push -u origin main
```

## 🧪 Test Locally First

```bash
cd /Users/sangmin/Developer/Qwen2.5-VL/huggingface-space
python test_local.py
```

## 🔑 API Key Required

Users will need a Dashscope API key from Alibaba Cloud:
- Get it here: https://bailian.console.alibabacloud.com/
- The key is entered by users, not stored in the app

## 📊 Features

- ✅ Text detection with bounding boxes
- ✅ Support for CJK languages (Chinese, Japanese, Korean)
- ✅ JSON output with coordinates
- ✅ Multiple model options (2B, 7B, 72B)
- ✅ Clean, user-friendly interface

## 🔧 Customization

To modify the app:
1. Edit `app.py` with your changes
2. Test locally with `python test_local.py`
3. Push updates:
   ```bash
   git add .
   git commit -m "Update: your changes"
   git push
   ```

## 📝 Notes

- The app uses Alibaba Cloud's API (no local model loading)
- Perfect for Hugging Face's free tier
- Auto-builds on push
- Public or private Space options available

## 🌐 After Deployment

Your Space URL will be:
```
https://huggingface.co/spaces/YOUR_USERNAME/qwen-ocr-detection
```

Share this URL with others to use your OCR app!

---

**Ready to deploy!** Choose your preferred method above and your OCR app will be live on Hugging Face Spaces in minutes. 🎉