https://www.unremarkable.ai/serverless-ai-inference-with-gemma-2-using-mozillas-llamafile-on-aws-lambda

We assume you have the following configured or installed.

1. An AWS account with credentials configured.
2. The AWS SAM CLI installed for fast and easy serverless deployments.
3. Docker installed for easy container builds and deployments.

After you clone the repo, setup your dependencies with the following command:

npm install

Now you can run the following commands from the root directory.

• ./bin/build - To download and build a llamafile container for deployment.
• ./bin/server - To run the download (above) llamafile server locally.
• ./bin/deploy - Deploy to AWS Lambda. Also does a build if needed.

This project uses Inquirer.js to chat with the model using OpenAI's API. The model can be running locally using bin/server or deployed to Lambda using bin/deploy. Inquirer will ask for your local or function URL at the beginning of the chat session.