This repo contains the code for the workshop on Generative AI Large Language Model Workshop for Financial Services. The workshop is designed to help you understand how to use leverage SageMaker to train, tune, and deploy Large Language Models.
If running this as part of an AWS hosted event, follow the instructions here to setup your environment.
If running this on your own, follow the instructions below to setup your environment.
- Make sure you have access to a SageMaker Studio environment. You can also use a SageMaker Notebook Instance or any other Jupyter Notebook environment that has programmatic access to AWS resources.
- Ensure your execution role has the following permissions:
SageMaker
CreateModel
CreateEndpointConfig / DeleteEndpointConfig
CreateEndpoint / DeleteEndpoint
CreateTrainingJob
SageMaker Runtime
InvokeEndpoint
- Clone this repo to your environment
git clone https://github.com/aws-samples/large-model-workshop-financial-services.git
cd large-model-workshop-financial-services
- Navigate to the
lab1
directory and open thefew_shot_learning.ipynb
notebook. Follow the instructions in the notebook to complete the lab.
Lab 1: Few Shot Learning - Introductory example showing how to fine-tune a sentence-transformer model for a classification task.
Lab 2: Large Language Model Tuning - Shows how to fine-tune a FLAN-T5 model for dialogue summarization.
Lab 3: Cost Effective Multi-Model Deployments - Shows how to deploy multiple models in a single endpoint to reduce inference costs.
See CONTRIBUTING for more information.
This repo is licensed under the MIT-0 License. See the LICENSE file.