About

Moondream is a lightweight, open-source Vision Language Model (VLM) designed to answer image-related questions, identify objects, and generate captions. With a compact 1.6 billion parameters, Moondream is highly efficient and can run on a variety of devices, including mobile phones and edge devices.