README.md

metadata

title: Grammar Correction App
emoji: 📝
colorFrom: pink
colorTo: purple
sdk: gradio
sdk_version: 4.14.0
app_file: app.py
pinned: false

📝 Grammar Correction App

A web-based grammar correction application built with Gradio and powered by a fine-tuned T5 transformer model. This application provides an intuitive interface for correcting grammatical errors in English text.

✨ Features

• Real-time Grammar Correction: Instantly fix grammatical errors in your text
• User-friendly Interface: Clean and intuitive web interface built with Gradio
• AI-Powered: Uses the vennify/t5-base-grammar-correction model from Hugging Face
• Example Sentences: Pre-loaded examples to demonstrate functionality
• Error Handling: Robust error handling for edge cases

🚀 Quick Start

Prerequisites

• Python 3.7 or higher
• pip package manager

Installation

1. Clone or download the repository

   git clone <repository-url>
   cd updated-grammar
   

2. Install dependencies

   pip install -r requirements.txt
   

3. Run the application

   python app.py
   

4. Access the web interface

   • Open your browser and navigate to the URL displayed in the terminal (typically http://127.0.0.1:7860)

📋 Dependencies

• transformers (4.36.2): Hugging Face library for transformer models
• torch (2.1.2): PyTorch deep learning framework
• gradio (4.8.0): Web interface framework for machine learning models

🛠️ How It Works

1. Model Loading: The app loads the pre-trained T5-base grammar correction model from Hugging Face
2. Text Processing: Input text is tokenized and formatted with a "grammar:" prompt
3. Inference: The model generates corrected text using beam search decoding
4. Output: The corrected text is displayed in the web interface

📝 Usage Examples

Try these examples in the application:

• Input: "She go to school every day." Output: "She goes to school every day."

• Input: "I is a boy." Output: "I am a boy."

• Input: "He don't like apples." Output: "He doesn't like apples."

• Input: "We was playing outside." Output: "We were playing outside."

⚙️ Model Configuration

The application uses the following model parameters:

• Model: vennify/t5-base-grammar-correction
• Max Input Length: 512 tokens
• Max Output Length: 128 tokens
• Beam Search: 5 beams
• Temperature: 0.7
• Early Stopping: Enabled

🔧 Customization

Modifying Model Parameters

You can adjust the generation parameters in the correct_grammar() function:

outputs = model.generate(
    **inputs,
    max_length=128,        # Maximum output length
    num_beams=5,          # Number of beams for beam search
    early_stopping=True,   # Stop when EOS token is generated
    temperature=0.7,       # Sampling temperature
    do_sample=False       # Use deterministic generation
)

Changing the Model

To use a different grammar correction model, update the model_path variable:

model_path = "your-preferred-model-name"

🚨 Error Handling

The application includes error handling for:

• Empty input text
• Model inference errors
• Tokenization issues

📱 Interface Features

• Input Box: Multi-line text area for entering text to correct
• Output Box: Displays the corrected text
• Examples: Click on example sentences to test the application
• Real-time Processing: Instant correction when you submit text

🤝 Contributing

Feel free to contribute to this project by:

1. Reporting bugs
2. Suggesting new features
3. Submitting pull requests
4. Improving documentation

📄 License

This project is open source. Please check the license file for more details.

🙏 Acknowledgments

• Hugging Face for providing the transformer models and libraries
• Gradio for the excellent web interface framework
• vennify for the pre-trained grammar correction model

🔗 Links

• Hugging Face Model
• Gradio Documentation
• Transformers Library

---

Note: The first run may take some time as the model needs to be downloaded from Hugging Face. Subsequent runs will be faster as the model will be cached locally.