LFM2-MoE: An 8.3B Parameter Language Model Running Entirely in Your Browser via WebGPU

Published by

on •1 min read

LFM2-MoE: An 8.3B Parameter Language Model Running Entirely in Your Browser via WebGPU

0:00 /0:57 1× A browser-based chat interface generating text from Liquid AI's LFM2-MoE model, running 8.3 billion parameters locally via WebGPU with no server backend. An 8.3-billion-parameter language model is running in your browser tab right now.

Showcase WebGPU MoE (Mixture of Experts)Transformers.js In-Browser Inference Liquid AI Hugging Face ONNX SEC

Learn more