Instructions

Installation

Note: This repository contains installation instructions for the MyGPT with Docker images and will not need source code. If you need access to MyGPT source to help in development process, please contact Jaimin Patel (Email: jaimin.patel@stjude.org) or appropriate person.

Note2: This repository is for public use and will not contain any private information. If you need access to private repository, please contact Jaimin Patel

MyGPT can be installed on following environments:

• Personal Computer
• Server/VM with GPU
• Cloud services (Azure)

Personal Computer

MyGPT is using Ollama for LLM server, and it requires at least 8GB (16GB for better response time) of RAM and 10GB of disk space. Also, Ollama is providing direct installation on Mac and Linux only. For Windows users we will use Docker to run Ollama.

To run the pipleine on following environments, follow the instructions:

• Mac

  • Basic Installation

• Linux

  • Basic Installation

• Windows

  • Basic Installation

  These instructions are simple and easy to follow. You can also modify bash scripts as per your convenience.

Server or VM with GPU

MyGPT can be hosted on a server or VM with GPU. For this installation we recommand to host User interface (UI), Backend server and Ollama (LLM server) on 3 seperate VMs. The Ollama VM should have a GPU with CUDA installed on the server/VM.

To run the pipleine on VM/Server, follow the instructions:

• Linux Server Installation

Cloud services (Azure)

MyGPT can be hosted on any cloud service but we are providing Azure as an example deploymnet. For this installation we recommand to host User interface (UI), Backend server and Ollama (LLM server) on 3 seperate VMs. The Ollama VM should have a GPU with CUDA installed on the VM.

To run the pipleine on Azure, follow the instructions:

• Azure Installation