facebook/sapiens2-normal-0.4b

Depth Estimation·facebook· 517· 3

sapiens2 other Depth Estimation 453.4M params arxiv:2604.21681base_model:facebook/sapiens2-pretrain-0.4bbase_model:finetune:facebook/sapiens2-pretrain-0.4blicense:otherregion:us

Per-pixel surface-normal estimation (3-channel unit vectors in camera frame).

Open in MLForge Sign up free Desktop app Source ↗

# pull & run locally
pip install mlforge-sdk && mlforge pull facebook/sapiens2-normal-0.4b

Model details

Task

Depth Estimation

Provider

facebook

Framework

sapiens2

Parameters

453.4M

Size

1.7 GB

License

other

Downloads

517

Likes

Paper

arXiv:2604.21681

Updated

2026-06-15

About facebook/sapiens2-normal-0.4b

Per-pixel surface-normal estimation (3-channel unit vectors in camera frame).

Related Depth Estimation

D depth-anything/Depth-Anything-V2-Small-hf Depth Estimation ·24.8M params 1.7M 38 🤗 HF D depth-anything/DA3METRIC-LARGE Depth Estimation 771.6K 19 🤗 HF D LiheYoung/depth-anything-large-hf Depth Estimation ·335.3M params 388.9K 65 🤗 HF D Intel/dpt-hybrid-midas Depth Estimation 225.1K 112 🤗 HF D depth-anything/DA3NESTED-GIANT-LARGE-1.1 Depth Estimation ·1.7B params 199.9K 27 🤗 HF D depth-anything/Depth-Anything-V2-Large-hf Depth Estimation ·335.3M params 199.1K 32 🤗 HF

Browse all Depth Estimation models →