---
library_name: pytorch
---

![moondream2_logo](resource/Moondream2.png)

Moondream is a highly efficient open-source vision language model that combines powerful image understanding capabilities with a remarkably small footprint, enabling efficient multimodal understanding with low latency.

- **Original paper / reference:** Moondream project repository — https://github.com/vikhyat/moondream

#  Moondream2-2B

This model implements the **Moondream2-2B**, optimized for efficient multimodal reasoning while maintaining a small compute footprint. It is well suited for applications such as visual question answering, image captioning, document understanding, and real-time multimodal assistants on edge or resource-constrained devices.

Model Configuration:
- Reference implementation: [moondream](https://github.com/vikhyat/moondream)
- Original Weight: [ moondream2](https://huggingface.co/vikhyatk/moondream2)
- Resolution: 3x378x378
- Support Cooper version:
    - Cooper SDK: [2.5.3]
    - Cooper Foundry: [2.2]

| Model | Device | Model Link |
| :-----: | :-----: | :-----: |
| Moondream2-2B | N1-655 | [Model_Link](https://huggingface.co/Ambarella/Moondream2/blob/main/n1-655_moondream2_2B.tar.gz) |
| Moondream2-2B | CV7 | [Model_Link](https://huggingface.co/Ambarella/Moondream2/blob/main/cv7_moondream2_2B.tar.gz) |
| Moondream2-2B | CV72 | [Model_Link](https://huggingface.co/Ambarella/Moondream2/blob/main/cv72_moondream2_2B.tar.gz) |