Quickly run an OpenAI-compatible chat completions server, using RWKV language models (such as EagleX). Self-contained no install dependencies required!
This guide is specifically for Windows users. minmodmon works on other platforms too, but no pre-compiled binaries are available currently.
- Download minmodmon from the latest stable release, or from the latest build.
- Unzip the archive, and run "minmodmon-server.exe".
- Open http://localhost:5000/ in your web browser, to open the dashboard.
- Download the model you want to use from the download link in the dashboard.
- Place the ".st" model file in the "data" directory.
- Re-start "minmodmon-server.exe".
- Under "Load Model", select the ID of the model you downloaded.
- Press "Load". Refresh the page to check loading status, until "Loaded model" changes to the model ID.
- In the top menu bar, press the "API" button.
- Select "Chat Completion" API type.
- Under "Chat Completion Source", select "Custom (OpenAI-compatible)"
- Under "Custom Endpoint (Base URL)", enter "http://localhost:5000/api"
- Press "Connect".
- Under "Available Models", select your previously loaded model. If only "None" is available, follow the instructions above to load a model.
Uses web-rwkv as the inference backend.