A Rough IdeaLOCAL MODEL PLAYGROUND
Text completions.
Write a prompt and see what comes next.
Not connected
Connect to the model running on this computer.
To switch models, open the playground on this computer: 127.0.0.1:8080.
LLM Explorables (llmexplorables.aroughidea.com):
To use this model there, put this line in .env and restart the model proxy server:
Connection settings
First time? Set it up: Docker, vLLM and the model proxy server, step by step.
Run vLLM and the model proxy server on the computer where you open this page. Allow this site’s local-network permission when prompted.
Start the model proxy server with this site’s address:
Completion
Your continuation will appear here.
First-token candidates
An approximation: temperature, then top-k, then top-p, renormalised over these candidates only. The faint bar behind is the model’s raw probability.
Requests
This page → the model proxy server. No key: the page never has it.
The model proxy server → vLLM. It adds the key from .env, masked here.