Hackathon Challenge - Irancell Second Place Winner

Overview

This repository contains the code that secured the second place in the Irancell hackathon challenge. The challenge involved developing a model capable of answering questions based on a pre-defined dataset and embeddings. The model includes a user interface (UI) built with the Streamlit library.

Requirements

Make sure you have the following libraries installed before running the code:

• pandas
• torch
• sentence-transformers

You can install them using the following command:

pip install pandas torch sentence-transformers

Code Structure

The main code file (trained_model.py) imports the necessary libraries and defines a function (answer_to_question) to generate responses based on the pre-defined dataset and embeddings.

• dataset_english.csv: Contains questions and answers for training the model.
• embeddings_english.csv: Stores embedding vectors used by the model.
• trained_model.py: The main script containing the code for the question-answering model.

Usage

1. Clone the repository:

   git clone https://github.com/77asadian/irancell_chatbot.git

2. Navigate to the project directory:

   cd project

3. Run the code:

   python streamlit run app.py

   The code initializes the model, processes user input, and provides answers based on the pre-trained embeddings.

Model Details

The model uses the "sentence-transformers/all-MiniLM-L6-v2" model for question encoding and semantic search. It filters and selects proper answers based on similarity scores.

Results

The model achieved second place in the Irancell hackathon challenge by effectively answering questions from the given dataset.

Contact

For any inquiries or collaborations, feel free to contact the repository owner:

• Hamidreza Asadian
• 77asadian@gmail.com
• My Linkedin)https://www.linkedin.com/in/77asadian/