LFM3-VL-2B-GGUF

Overview

The vision member of the LFM3 family at Q4_K_M - image understanding on top of the same agent-trained tri-hybrid backbone (load the mmproj file alongside the model). Edge vision agents with a 256K context window.

Good for

• agentic

Recommended hardware

≈ 4 GB VRAM to run this 2B model at the Q4_K_M quantization.

Fits 4 GB+ consumer GPUs (e.g. GTX 1650), or CPU inference

Links

• View on HuggingFace
• Run on Infersec