aws-samples/fine-tuning-llm-with-domain-knowledge

This repo walks you through how to use transfer learning to fine tune a LLM (large language model) using UK Supreme Court case law as the domain specific training dataset. The model being fine-tuned is the HuggingFace GPTJ-6B model.

37
/ 100
Emerging

This project helps machine learning engineers or data scientists enhance a large language model's ability to understand and generate text specific to a particular domain. By feeding it a dataset of UK Supreme Court case law, you can transform a general-purpose language model into one proficient in legal terminology and context. The input is a pre-trained LLM and domain-specific text documents, and the output is a fine-tuned LLM ready for specialized tasks.

No commits in the last 6 months.

Use this if you need to create a language model that is highly knowledgeable and accurate within a very specific field, such as legal research or medical documentation, beyond what a general LLM can offer.

Not ideal if you are a legal professional looking for a direct legal advice tool, or if you lack experience with AWS SageMaker and machine learning model training.

machine-learning-engineering natural-language-processing legal-tech model-customization data-science
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 8 / 25
Maturity 16 / 25
Community 13 / 25

How are scores calculated?

Stars

42

Forks

6

Language

Jupyter Notebook

License

MIT-0

Category

llm-fine-tuning

Last pushed

Aug 08, 2023

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/transformers/aws-samples/fine-tuning-llm-with-domain-knowledge"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.