README.md

---
library_name: pytorch
---

![depthanythingv2_logo](resource/DepthAnythingV2.png)

DepthAnythingV2 is a lightweight depth estimation model designed to predict accurate per-pixel depth maps from a single RGB image, optimized for efficiency and versatility across scenes.

Original paper: [DepthAnything V2](https://arxiv.org/abs/2406.09414)

# DepthAnythingV2-Small

This model uses the DepthAnythingV2-Small variant, which balances model size and inference speed while maintaining strong depth estimation accuracy. It is well suited for applications such as AR/VR, robotics, scene reconstruction, and real-time 3D perception on edge devices.

Model Configuration:
- Reference implementation: [Official DepthAnythingV2 source code](https://github.com/DepthAnything/Depth-Anything-V2)
- Original Weight:[Depth-Anything-V2-Small](https://huggingface.co/depth-anything/Depth-Anything-V2-Small/resolve/main/depth_anything_v2_vits.pth?download=true)
- Resolution: 3x224x224
- Support Cooper version:
    - Cooper SDK: [2.5.2]
    - Cooper Foundry: [2.2]

| Model | Device | Model Link |
| :-----: | :-----: | :-----: |
| DepthAnythingV2-Small | N1-655 | [Model_Link](https://huggingface.co/Ambarella/DepthAnythingV2/blob/main/n1-655_depthanything_v2_small.bin) |
| DepthAnythingV2-Small | CV72 | [Model_Link](https://huggingface.co/Ambarella/DepthAnythingV2/blob/main/cv72_depthanything_v2_small.bin) |
| DepthAnythingV2-Small | CV75 | [Model_Link](https://huggingface.co/Ambarella/DepthAnythingV2/blob/main/cv75_depthanything_v2_small.bin) |