Introducing LFM2-VL — our new generation of efficient vision-language models for real-world deployment, from smartphones and laptops to wearables and embedded systems. 🧵