LHM 500M HF
LHM is a feed-forward model capable of reconstructing animatable 3D human bodies from a single image in seconds.
Downloads 456
Release Time : 3/28/2025
Model Overview
Trained with image reconstruction loss on large-scale video datasets, LHM demonstrates strong generalization capabilities across diverse real-world scenarios, enabling rapid generation of animatable 3D human models from a single image.
Model Features
Second-level 3D Reconstruction
Capable of generating animatable 3D human models from a single image in seconds.
Large-scale Training
Trained with image reconstruction loss on large-scale video datasets, exhibiting strong generalization capabilities.
Diverse Scenario Adaptation
Able to process image inputs from diverse real-world scenarios.
Model Capabilities
Single-image 3D Reconstruction
Animatable Human Generation
Fast Inference
Use Cases
Digital Entertainment
Game Character Creation
Quickly generate game character models from player-uploaded photos
Reduces character creation time and enhances user experience
Virtual Reality
Virtual Avatar Generation
Rapidly create user virtual avatars for VR applications
Lowers the barrier to virtual avatar creation
Featured Recommended AI Models
Š 2025AIbase