Extended and Ongoing Notes: Amazon Web Services Fine-Tunes a Large Language Model
(On-going)
I started to experiment with fine-tuning Open-weights Large Language Models (LLMs). I'm using Amazon Webservices (AWS) SageMaker. In the following Medium article, I describe a basic checklist I used to get up to speed (AWS is a complicated - and powerful platform).
I'm learning more as I progress, do more, and become more involved in the AWS system and services. Below is a running list of resources, discoveries, and thoughts I found helpful moving forward.
This post is a living document - I'll update this page as I collect insight. In the future - when I've collected enough, I may roll the contents of this post into an updated Medium article. I’ll send out an email when I do.
New References:
AWS > Documentation > Amazon SageMaker > Developer Guide > Clean Up - Despite being careful, I incurred unexpected costs with my AWS. Rigor with hygiene is a good idea. This is the AWS checklist for cleaning up SageMaker.
Amazon SageMaker Studio—A Fully Integrated Development Environment For Machine—As the Medium article hints, I’m not using a remote command line interface (AWS CLI) to interact with AWS at this time. I am using Studio, which provides a deep toolset.