README.md

metadata

title: Intent Classifier Chatbot
emoji: 🤖
colorFrom: blue
colorTo: purple
sdk: docker
pinned: false
app_port: 7860
license: apache-2.0
short_description: Intent Detection API using BERT and Flask

🤖 Bert Intent Chatbot (Encoder-Only)

Intent Detection API using BERT and Flask
This project uses a fine-tuned BERT model that is an encoder only model without a decoder used to detect user intent from text inputs (e.g., chatbot queries). It provides a lightweight Flask API to classify input sentences into predefined intent categories.

---

🔧 Features

• ✅ Pretrained BERT fine-tuned on CLINC150
• 🧠 Real-time intent classification from natural text
• 🌐 REST API using Flask
• 🤖 Easy to integrate with chatbots, voice assistants, or NLP systems

---

📊 Dataset: CLINC150

The project uses the CLINC150 dataset, a benchmark dataset for intent classification in task-oriented dialogue systems.

🧾 Overview

• Total intents: 150 unique user intents
• Domains: 10 real-world domains (e.g., banking, travel, weather, small talk)
• Examples: ~22,500 utterances
• Language: English
• Out-of-scope (OOS): Includes OOS examples to test robustness

📁 Dataset Splits

Split	Examples
Train	7,000
Validation	3,000
Test	5,500

📦 Source

• Official repo: clinc/oos-eval
• Hugging Face: clinc_oos

---

🚀 Example

Request

{
  "text": "I want to book a flight"
}

Response

{
  "intent": "book_flight"
}

---

Intent Classifier Chatbot

A sophisticated intent classification system built with BERT and FastAPI that can predict user intents from natural language text.

Features

• Advanced NLP: Uses BERT-based transformer model for accurate intent classification
• 150+ Intent Classes: Trained on the CLINC150 dataset with comprehensive intent coverage
• Real-time Prediction: FastAPI backend for fast inference
• Clean UI: Simple and intuitive web interface
• Production Ready: Dockerized for easy deployment

How to Use

1. Enter your message in the text area
2. Click "Predict Intent"
3. See the AI's prediction of your intent

Try examples like:

• "Set an alarm for 7am" → Alarm
• "Transfer money to John" → Transfer
• "What's the weather like?" → Weather
• "Book a flight to Paris" → Book Flight

Model Details

• Architecture: BERT for Sequence Classification
• Dataset: CLINC150 (151 intent classes including out-of-scope)
• Accuracy: High performance on intent classification tasks
• Preprocessing: Advanced tokenization and text normalization

Tech Stack

• Backend: FastAPI, PyTorch, Transformers
• Frontend: HTML, CSS, JavaScript
• Model: BERT-base fine-tuned on CLINC150
• Deployment: Docker, Hugging Face Spaces

Author

Saher Muhamed

• GitHub: @sahermuhamed1
• Email: sahermuhamed176@gmail.com

🤗 Hugging Face Spaces Configuration

To deploy this project on Hugging Face Spaces, you can use a README.md and a config.json file to configure your Space for inference.
Check out the configuration reference at:
https://huggingface.co/docs/hub/spaces-config-reference

Example config.json for inference API:

{
  "hf_space_type": "inference",
  "python_version": "3.10",
  "sdk": "gradio",
  "entrypoint": "app.py"
}

• Make sure your requirements.txt lists all dependencies.
• The entrypoint should point to your main app file (e.g., app.py or main.py).
• For more details and advanced configuration, see the Spaces config reference.