kasinathansj/Categorization

Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

Text Classification API

This is a FastAPI-based service that classifies text into predefined departments using the Mistral-7B-Instruct model.

Features

• —Classifies input text into the most relevant department.
• —Uses the Mistral-7B-Instruct model for classification.
• —Returns structured JSON responses.
• —Provides an API endpoint for classification.
• —Logs execution time for performance analysis.

Space Configuration

This project is hosted on Hugging Face Spaces using Docker.

Setup & Installation

1. 1.Ensure the required dependencies are installed by specifying them in requirements.txt:

   fastapi
   uvicorn
   pydantic
   llama-cpp-python
   huggingface-hub
   ...

1. 1.Place the model file mistral-7b-instruct-v0.1.Q2_K.gguf in the project directory.

1. 1.Run the application using the following command:

   uvicorn app:app --host 0.0.0.0 --port 7860

API Endpoints

Base Url

https://kasinathansj-categorization.hf.space

Home

GET `/`

Response Body:

{
    "message": "Text classification API is running!"
}

Classify Text

POST `/classify`

Request Body:

{
    "text": "The IT department needs new software updates.",
    "departments": [
        {"name": "IT", "description": "Handles software and hardware issues."},
        {"name": "HR", "description": "Manages employee relations."}
    ]
}

Response Body:

{
    "text": "The IT department needs new software updates.",
    "department": "IT",
    "reason": "The text pertains to software and hardware issues.",
    "time_taken": 1.23
}

Logs

Application logs are visible in the Hugging Face Space runtime output.

License

MIT License

Author

Kasinathan SJ