NLP and LLM Learning Workspace

This repository contains a collection of Jupyter notebooks and resources for learning and experimenting with Natural Language Processing (NLP) and Large Language Models (LLMs). The project is organized into different subdirectories, each focusing on specific topics or tutorials.

Project Structure

.gitignore
README.md
notebooks/
    huggingface_learning/
        datasets_tutorial.ipynb
    nlp_learning/
        tokenization.ipynb
    SQLDB_chain/
        experiment_001.ipynb

Notebooks

1. Hugging Face Learning

   • datasets_tutorial.ipynb: Demonstrates how to use the Hugging Face datasets library to load and process datasets like WMT14 for machine translation tasks.

2. NLP Learning

   • tokenization.ipynb: Explores tokenization techniques using spaCy for English and French text, vocabulary building, and dataset preprocessing for translation tasks.

3. SQLDB Chain

   • experiment_001.ipynb: Experiments with the langchain_community library for SQL database tools, including error handling and debugging.

Requirements

• Python 3.11 or higher
• Jupyter Notebook
• Required Python libraries:
  • spacy
  • datasets
  • sqlalchemy
  • torch
  • langchain_community

Setup

1. Clone the repository:

   git clone <repository-url>
   cd NLP_LLM_Coding

2. Create a virtual environment:

   python -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install dependencies:

   pip install -r requirements.txt

4. Download spaCy language models:

   python -m spacy download en_core_web_sm
   python -m spacy download fr_core_news_sm

Usage

• Open the Jupyter notebooks in the notebooks/ directory to explore the tutorials and experiments:

  jupyter notebook

License

This project is licensed under the MIT License. See the LICENSE file for details.

Acknowledgments

• Hugging Face for the datasets library.
• spaCy for NLP tools.
• LangChain for SQL database tools.