Back to Apple
Research PreviewSpecializedimage3dUpdated March 14, 2025

LiTo

Model Overview

LiTo is a undisclosed-parameter model developed by Apple. Released on 2025-03-14.


📊 Quick Specs

Specification Table
SpecificationValue
Parametersundisclosed
Taskother
Modalityimage, 3d
LicenseOther/Custom
Typeresearch-preview

✨ Key Features

  • Unified 3D Latent Representation: Simultaneously encodes surface geometry and view-dependent appearance (surface light fields) into a compact latent code
  • Image-to-3D Generation: Single-image feed-forward reconstruction of full 3D objects using latent flow matching
  • Surface Light Field Tokenization: Tokenizes view-dependent radiance on the object surface for efficient and disentangled modeling
  • Latent Flow Matching: Uses flow matching in the LiTo latent space for high-quality, diverse 3D generation
  • ICLR 2026 Paper: Peer-reviewed research published at the International Conference on Learning Representations 2026

🔗 Resources


📜 License & Access

Other/Custom — See repository for specific license details.

Key Features

Unified 3D Latent Representation: Simultaneously encodes surface geometry and view-dependent appearance (surface light fields) into a compact latent code

Feature 01

Image-to-3D Generation: Single-image feed-forward reconstruction of full 3D objects using latent flow matching

Feature 02

Surface Light Field Tokenization: Tokenizes view-dependent radiance on the object surface for efficient and disentangled modeling

Feature 03

Latent Flow Matching: Uses flow matching in the LiTo latent space for high-quality, diverse 3D generation

Feature 04

ICLR 2026 Paper: Peer-reviewed research published at the International Conference on Learning Representations 2026

Feature 05

You might also want to compare

Verified Sources

Tags

3d-generationneural-renderingimage-to-3dsurface-light-fieldflow-matching

Model Specs

research-preview

Parameters

Undisclosed

Context Window

unknown

License

Other/Custom

Deployment

self-hostable

Resources & Links

Curator Notes

Partially enriched via migration on 2026-07-25. Manual review recommended.

Compare Specs

Compare parameters, context windows, modalities, and benchmark scores of this model side-by-side with others.

Compare Model