Spaces:

jeyanthangj2004
/

ocr

Runtime error

edocr2/
├── app.py                          # Gradio application
├── requirements.txt                # Python dependencies
├── packages.txt                    # System dependencies
├── README.md                       # Space description
├── .gitattributes                  # Git LFS configuration
├── edocr2/                         # eDOCr2 package
│   ├── __init__.py
│   ├── tools/
│   ├── keras_ocr/
│   └── models/                     # Model files (IMPORTANT!)
│       ├── recognizer_gdts.keras   # 67.2 MB
│       ├── recognizer_gdts.txt
│       ├── recognizer_dimensions_2.keras  # 67.2 MB
│       └── recognizer_dimensions_2.txt
└── tests/
    └── test_samples/               # Example images (optional)

🔧 Step-by-Step Deployment

Step 1: Download Model Files

Download the 4 model files from GitHub Releases:

cd edocr2-main
mkdir -p edocr2/models
cd edocr2/models

# Download models (Linux/Mac)
wget https://github.com/javvi51/edocr2/releases/download/v1.0.0/recognizer_gdts.keras
wget https://github.com/javvi51/edocr2/releases/download/v1.0.0/recognizer_gdts.txt
wget https://github.com/javvi51/edocr2/releases/download/v1.0.0/recognizer_dimensions_2.keras
wget https://github.com/javvi51/edocr2/releases/download/v1.0.0/recognizer_dimensions_2.txt

cd ../..

Windows PowerShell:

cd edocr2-main
New-Item -ItemType Directory -Force -Path edocr2\models
cd edocr2\models

Invoke-WebRequest -Uri "https://github.com/javvi51/edocr2/releases/download/v1.0.0/recognizer_gdts.keras" -OutFile "recognizer_gdts.keras"
Invoke-WebRequest -Uri "https://github.com/javvi51/edocr2/releases/download/v1.0.0/recognizer_gdts.txt" -OutFile "recognizer_gdts.txt"
Invoke-WebRequest -Uri "https://github.com/javvi51/edocr2/releases/download/v1.0.0/recognizer_dimensions_2.keras" -OutFile "recognizer_dimensions_2.keras"
Invoke-WebRequest -Uri "https://github.com/javvi51/edocr2/releases/download/v1.0.0/recognizer_dimensions_2.txt" -OutFile "recognizer_dimensions_2.txt"

cd ..\..

Step 2: Install Git LFS

Linux/Mac:

# Ubuntu/Debian
sudo apt-get install git-lfs

# macOS
brew install git-lfs

# Initialize
git lfs install

Windows: Download from git-lfs.github.com and install.

Step 3: Create Hugging Face Space

Go to huggingface.co/spaces
Click "Create new Space"
Fill in details:
- Space name: edocr2 (or your choice)
- License: MIT
- SDK: Gradio
- Hardware: CPU Basic (free) or GPU (paid)
Click "Create Space"

Step 4: Clone the Space Repository

# Clone your new space
git clone https://huggingface.co/spaces/YOUR_USERNAME/edocr2
cd edocr2

Step 5: Copy Files to Space

# Copy all files from edocr2-main to the space
cp -r /path/to/edocr2-main/* .

# Or on Windows
xcopy /E /I C:\path\to\edocr2-main\* .

Step 6: Track Large Files with Git LFS

# Track model files
git lfs track "*.keras"
git lfs track "edocr2/models/*.keras"

# Verify LFS is working
git lfs ls-files

Step 7: Commit and Push

# Add all files
git add .

# Commit
git commit -m "Initial deployment of eDOCr2"

# Push to Hugging Face
git push origin main

Note: Uploading large files (140 MB total) may take several minutes.

Step 8: Wait for Build

Go to your Space URL: https://huggingface.co/spaces/YOUR_USERNAME/edocr2
Wait for the build to complete (5-10 minutes)
Check the "Logs" tab for any errors

✅ Verification

Once deployed, your Space should:

✅ Show the Gradio interface
✅ Load models successfully (check logs)
✅ Accept image uploads
✅ Process drawings and show results

🐛 Troubleshooting

Build Fails

Error: No module named 'edocr2'

Solution: Ensure the edocr2/ folder is in the root directory.

Models Not Found

Error: Model files not found!

Solution:

Check edocr2/models/ contains all 4 files
Verify Git LFS tracked the .keras files
Re-push with git lfs push --all origin main

Out of Memory

Error: Killed or memory errors

Solution:

Upgrade to GPU hardware (Settings → Hardware)
Or reduce max_img_size in app.py (line with max_img_size=1048)

Timeout Errors

Error: Processing takes too long

Solution:

Upgrade to GPU hardware
Reduce image resolution before upload
Set timeout in Space settings

⚙️ Configuration

Change Hardware

Go to Space Settings
Select Hardware tab
Choose:
- CPU Basic (Free) - Slower, ~30s per image
- CPU Upgrade ($0.03/hr) - Faster
- T4 GPU ($0.60/hr) - Fastest, ~10s per image

Set as Private

Go to Space Settings
Toggle "Private" switch
Only you can access the Space

Add Secrets

For API keys or sensitive data:

Settings → Variables and secrets
Add New secret
Access in code: os.environ['SECRET_NAME']

📊 Monitoring

View Logs

Click "Logs" tab in your Space
Monitor model loading and processing
Check for errors

Analytics

Go to "Analytics" tab
View usage statistics
Monitor performance

🔄 Updating the Space

To update your deployed Space:

# Make changes to files
# Then commit and push
git add .
git commit -m "Update: description of changes"
git push origin main

The Space will automatically rebuild.

💰 Cost Estimation

Free Tier (CPU Basic)

✅ Unlimited usage
⏱️ ~20-30s per image
💾 2 GB RAM
🔋 2 vCPU

Paid Tier (T4 GPU)

💵 $0.60/hour
⏱️ ~5-10s per image
💾 16 GB RAM
🚀 Much faster

Tip: Use CPU Basic for testing, upgrade to GPU for production.

🌐 Custom Domain

To use a custom domain:

Go to Space Settings
"Custom domain" section
Add your domain
Follow DNS configuration instructions

📱 Embedding

Embed your Space in a website:

<iframe
  src="https://YOUR_USERNAME-edocr2.hf.space"
  frameborder="0"
  width="850"
  height="450"
></iframe>

🔗 Useful Links

Hugging Face Docs: huggingface.co/docs/hub/spaces
Gradio Docs: gradio.app/docs
Git LFS: git-lfs.github.com

🎉 Success!

Your eDOCr2 Space is now live! Share it with:

https://huggingface.co/spaces/YOUR_USERNAME/edocr2

Need help? Open an issue on GitHub