# One-2-3-45/One-2-3-45

[NeurIPS 2023] Official code of "One-2-3-45: Any Single Image to 3D Mesh in 45 Seconds without Per-Shape Optimization"

Repository: https://github.com/One-2-3-45/One-2-3-45
Canonical: https://ross.abutalabs.com/products/one-2-3-45
Homepage: http://one-2-3-45.com
Language: Python
License: Apache-2.0
License Family: permissive
Topics: 3d, 3d-generation, aigc, generalizable-nerf, mesh-generation, multi-view-reconstruction, nerf, 3d-shape-generation
Last push: 2024-04-20T23:15:25+00:00

## Health v2 (maintenance only)
Score: 29/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 0, release rhythm 35, longevity 83
- inputs: {"age_days": 1162, "days_push": 865, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1718, forks 103 (observed 2026-08-28T04:05:26.553556+00:00)

## What it is
One-2-3-45 is the official PyTorch implementation of a NeurIPS 2023 paper that converts any single image into a full 360-degree 3D textured mesh in about 45 seconds. It uses the Zero123 view-conditioned diffusion model to generate multi-view images and an SDF-based generalizable neural surface reconstruction module to lift them into 3D, avoiding slow per-shape optimization.

## Use cases
- generate a 3D mesh from a single photo
- reconstruct 3D models from images without per-shape optimization
- convert text prompts into 3D meshes via text-to-image diffusion
- create multi-view images of an object from one view
- research on generalizable neural surface reconstruction
- run image-to-3D inference via provided APIs or Hugging Face demo

## When to choose
- you need fast single-image to 3D mesh generation
- you want better geometry and 3D consistency than SDS-based optimization methods
- you have an NVIDIA GPU with at least 18GB memory
- you want to extend or study feed-forward 3D reconstruction from multi-view diffusion outputs

## When to avoid
- you need production-grade 3D asset pipelines with fine topology control
- you lack a CUDA GPU with sufficient VRAM
- you need real-time or mobile 3D reconstruction
- you want actively maintained software beyond the paper's lifecycle

## Facets
- artifact type: library
- maturity: stable
- function: machine-learning, deep-learning, image-processing, graphics, simulation
- domain: computer-vision, graphics, artificial-intelligence, deep-learning
- platform: python
- tags: 3d-generation, mesh-generation, nerf, multi-view-reconstruction, diffusion-models, text-to-3d, image-to-3d, aigc, research-code, linux, gpu

## Member repositories
- One-2-3-45/One-2-3-45 (main) score 29

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:26.553556+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T03:34:09.392561+00:00, confidence not recorded.
  - readme: https://github.com/One-2-3-45/One-2-3-45 (fetched 2026-08-28T04:05:26.553556+00:00, sha 8e4c71bc720c)
  - homepage: http://one-2-3-45.com (fetched 2026-08-29T11:10:06.424112+00:00, sha df95e8cbdda2)
- Data as of 2026-08-30T08:39:29.467469+00:00.