/train_scheduling_assistant

This project utilizes a fine-tuned Large Language Model (LLM) to generate train scheduling information from unstructured textual data, providing an interactive UI via Streamlit.

Primary LanguagePython

Train Scheduling with Large Language Models

This project utilizes a fine-tuned Large Language Model (LLM) to generate train scheduling information from unstructured textual data, providing an interactive UI via Streamlit.

Demo

Below is a short demo of the project in action:

Demo

Table of Contents

Installation and Setup

git clone https://github.com/fshnkarimi/train_scheduling_assistant.git
cd train_scheduling_assistant
pip install -r requirements.txt

Usage

Run the Streamlit app locally:

streamlit run app.py

Navigate to http://localhost:8501 in your web browser to interact with the application.

Project Structure

  • llm/: Contains files related to the fine-tuning and usage of the LLM.
  • data/: Store your synthetic and real-world data for training and evaluation.
  • nlp/: Contains Natural Language Processing utilities for text preprocessing and information extraction.
  • models/: Place to store the fine-tuned LLM model.
  • app.py: Streamlit application for user interaction and visualization.
  • requirements.txt: List of Python dependencies required for the project.

Technologies Used

  • Python
  • PyTorch
  • Hugging Face Transformers
  • Streamlit
  • Docker
  • Kubernetes