/LLM

A package demonstrating LLM implementation and practical deployment

Primary LanguageJupyter NotebookGNU General Public License v3.0GPL-3.0

LLM

A package demonstrating model implementation and deployment for solving practical problems.

Getting Started

Get the package

# Navigate to your local folder
cd /your/local/folder

# Clone the WindML repository
git clone git@github.com:marcodigennaro/LLM.git

# Enter the folder
cd LLM/

# Create the python environment from the pyproject.toml file
poetry install

# Activate the python environment
source .venv/bin/activate

# Run tests 
poetry run pytest -v

# Start Jupyter Lab
jupyter-lab

Content of the Jupyter Notebooks

  1. sentiment_analysis shows three strategies for sentiment analysis on a database of amazon reviews

    1. VADER (bag of words)
    2. RoBERTa (Transformer)
    3. Transformers Pipelines

  2. RNN demonstrate how to use a basic RNN as LLM

    1. Reads a text (Prometheus's) legend and predict the continuation of the story
    2. Uses MDS, Isomaps and t-SNE representations to project the results onto a 2D space

VADER vs RoBERTa Accuracy Prometheus

Author

License

This project is licensed under the GPL v3 License - see the LICENSE.md file for details

Acknowledgments