Skip to content

A comprehensive guide and codebase for text summarization using Large Language Models (LLMs). Dive into techniques, from chunking to clustering, and harness the power of LLMs like GPT-3.5 and GPT-4.

Notifications You must be signed in to change notification settings

datamade/llm-text-summarization

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

14 Commits
 
 
 
 
 
 
 
 

Repository files navigation

LLM-based Text Summarization: Novice to Maestro 🚀

A comprehensive guide and codebase for text summarization harnessing the capabilities of Large Language Models (LLMs). Delve deep into techniques, from chunking to clustering, and maximize the potential of LLMs like GPT-3.5 and GPT-4.

📄 Article: I highly recommend reading this article before diving into the code.

Architecture

Full_Architecture

🚀 Getting Started

  1. Clone the Repository
  2. Install from setup.py:
    python3 -m pip install .
  3. Install Spacy's English Dataset:
    python3 -m spacy download en_core_web_sm
  4. Set Up OpenAI API Key:
    export OPENAI_API_KEY='sk-...'
  5. Configure settings: Copy the example settings file and modify as needed.
    cp config.example.yaml config.yaml
  6. File Handling: For the input file, only .txt is accepted. For the output, .json is preferred.
  7. Run the Program:
    llm-summarize my_input.txt my_output.json

🛠 Understanding the config.yaml File

  • summary_type_token_limit: Determines how to categorize the input text: short, medium, or long.
  • sentence_splitter: Adjust approx_total_doc_tokens. Keep it around 1000 for medium-sized texts and up to 6000 for longer texts.
  • cod: Configuration for Chain of Density (CoD) prompting.
  • map_reduce: To further condense the final summary with CoD, set final_dense to true.
  • cluster_summarization: Adjust num_closest_points_per_cluster (max value: 3) for the top-k best chunks. Vary num_clusters (hyper-parameter for k-means) to optimize results.
  • Remaining configs are self-explanatory.

📋 Output JSON Format

The output JSON comprises:

{
    "summary": "Descriptive final summary...",
    "keywords": ["Keyword1", "Keyword2", "..."],
    "metadata": {
        "total_tokens": 3625,
        "total_cost": 0.082,
        "total_time": 86.23
    }
}
  • summary: The final summary output
  • keywords: important keywords and phrases
  • metadata: Provides total time (in seconds) taken to execute your summary, total cost (in USD) for openai, and total token counts in the whole process

❤️ If this repository helps, please star ⭐, and share ✔️!
If you also found the article informative and think it could be beneficial to others, I'd be grateful if you could like 👍, follow 👉, and share✔️ the piece with others.
Happy coding!

About

A comprehensive guide and codebase for text summarization using Large Language Models (LLMs). Dive into techniques, from chunking to clustering, and harness the power of LLMs like GPT-3.5 and GPT-4.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 100.0%