D

Dpt Beit Large 512

Developed by Intel
A monocular depth estimation model based on BEiT Transformer, capable of inferring fine depth information from a single image
Downloads 2,794
Release Time : 11/28/2023

Model Overview

This DPT model uses the BEiT model as its backbone network, with added neck and head structures for monocular depth estimation, applicable in fields such as generative AI, 3D reconstruction, and autonomous driving.

Model Features

High-quality depth estimation
Utilizes BEiT Transformer to achieve the highest quality depth estimation results
Multi-resolution support
Offers variants like BEiT512-L, BEiT384-L, and BEiT384-B, supporting different training resolutions
Zero-shot transfer capability
Features zero-shot transfer capability with a metric value of 10.82

Model Capabilities

Monocular depth estimation
Image depth information inference
Zero-shot transfer

Use Cases

Computer vision
3D reconstruction
Infers depth information from a single image for 3D scene reconstruction
Autonomous driving
Provides environmental depth perception for autonomous driving systems
Generative AI
Supplies depth information as input for generative AI models
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
Š 2025AIbase