Depth-Anything-V2-Small-hf
depth-anything/Depth-Anything-V2-Small-hf
A popular open depth estimation model, with 1.7M downloads a month. gigarouter benchmarks and hosts it as an OpenAI-compatible API.
about this model
Depth-Anything-V2-Small-hf is a monocular depth estimation model that predicts dense depth from a single image. It uses a DPT architecture with a DINOv2 backbone and is trained on approximately 600,000 synthetic labeled images and over 62 million real unlabeled images.
Key strengths
- More fine-grained detail than Depth Anything V1
- More robust predictions than Depth Anything V1 and diffusion-based models such as Marigold and Geowizard
- Roughly 10× faster and significantly more lightweight than diffusion-based alternatives
- Pre-trained checkpoints that yield strong fine-tuned performance
Intended use
This model is suited for zero-shot depth estimation across diverse scenes. It achieves state-of-the-art results for both relative and absolute depth estimation.
Benchmark results
The model card reports state-of-the-art performance on standard depth estimation benchmarks, but does not provide specific numeric values. For further details, refer to the paper Depth Anything V2.
Additional resources
We're benchmarking and onboarding Depth-Anything-V2-Small-hf as a hosted, OpenAI-compatible API. Sign in for free credit and be ready when it lands, or tell us you want it and we'll prioritize it.