search

cpacker / MemGPT

Create LLM agents with long-term memory and custom tools 📚🦙
https://memgpt.readme.io
Apache License 2.0
10.79k stars 1.16k forks source link

deployment with docker compose #1430

Open wwjCMP opened 1 month ago

wwjCMP commented 1 month ago

How do I set up my deployment with docker compose to use the ollama model and ollama Embed model

sarahwooders commented 1 month ago

We've only tested with OpenAI so far, so not sure if it'll work with Ollama (you can do memgpt server and that should work with Ollama though). We also are adding embedding support in #1384.

It's a bit hacky right now, but the following might also work

  1. Install pip install pymemgpt
  2. Run memgpt configure and configure Ollama settings
  3. Run docker compose up and change this line to point to your file ~/.memgpt/config