LFM2 is a new generation of hybrid models, designed for on-device deployment.
AI & ML interests
A new generation of foundation models from first principles.
Recent Activity
spaces 10
Running
19
LFM2 MoE WebGPU
💧
WebGPU-accelerated Mixture of Experts
Configuration error
Featured
87
LFM2 WebGPU – In-browser tool calling
🛠
In-browser tool calling, powered by Transformers.js
Configuration error
11
LFM2.5-Audio WebGPU Demo
🎙
ASR, TTS & conversational audio in your browser
Running
Featured
54
LFM2 VL WebGPU
📹
Real-time video captioning in your browser
Build error
Featured
36
LFM2 MCP
🚀
In-browser tool calling with MCP, powered by Transformers.js
Running
37
Wllama
🦙
Run GGUF directly on your browser!
models 93
LiquidAI/LFM2.5-350M
Text Generation • 0.4B • Updated • 2.99k • 57
LiquidAI/LFM2.5-350M-ONNX
Text Generation • Updated • 4 • 4
LiquidAI/LFM2-2.6B-Transcript
Text Generation • 3B • Updated • 271 • 161
LiquidAI/LFM2-1.2B-Tool
Text Generation • 1B • Updated • 345 • 95
LiquidAI/LFM2-350M-Math
Text Generation • 0.4B • Updated • 513 • 55
LiquidAI/LFM2.5-350M-Base
Text Generation • 0.4B • Updated • 3
LiquidAI/LFM2.5-350M-MLX-bf16
Text Generation • 0.4B • Updated • 4
LiquidAI/LFM2.5-350M-MLX-4bit
Text Generation • 55.4M • Updated • 2 • 2
LiquidAI/LFM2.5-350M-MLX-5bit
Text Generation • 66.5M • Updated • 2 • 2
LiquidAI/LFM2.5-350M-MLX-6bit
Text Generation • 77.6M • Updated • 1 • 2