Prerequisites

Running all containers on your device

You’ll need docker desktop installed which you can download from their website.

You can use curl from command line as shown in the installation verification below or you can use Postman which is one of many good tools for calling web services. You can download Postman from their website.

With all the services running on your local machine you wont be able to load much data or put much demand on it. You should have at least 32G ram, but 64G ram would be good. The system may work on less ram but has not been tested. It can detect CUDA and Apple Silicon GPUs for accelerated usage of pytorch.

Running web service or queue worker directly

  • Python3 (v3.11+ recommended)

  • Pip

  • Make (a cli build system) - Usually installed on linux systems already

If you want to debug through the service or worker using our pre-made configurations:

  • VSCode (a developer ide)

Running all container in AWS EKS Cluster

  • AWS Account with privileges to run CloudFormation, create and destroy resources, etc.

  • AWS CLI installed with profile configuration

  • AWS CDK installed

This project comes with an AWS CDK application which will setup a new VPC with EKS cluster running the containers from this project.

WARNING: The generic installation of the EKS cluster is on 2 EC2 D-Large servers that each cost 8 cents per hour they are on. That is 16 cents per hour for both, which is approximately $115 per month if run every hour of the month. That does not cost other fees associated with using the system such as network traffic costs.