# aigc3d/LHM

[ICCV2025] LHM: Large Animatable Human Reconstruction Model from a Single Image in Seconds

Repository: https://github.com/aigc3d/LHM
Canonical: https://ross.abutalabs.com/products/lhm
Homepage: https://aigc3d.github.io/projects/LHM/
Language: Python
License: Apache-2.0
License Family: permissive
Topics: aicg, aigc, digitalhuman
Last push: 2026-03-17T05:17:23+00:00

## Health v2 (maintenance only)
Score: 52/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 72, release rhythm 35, longevity 38
- inputs: {"age_days": 538, "days_push": 169, "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 2664, forks 213 (observed 2026-08-28T04:07:08.968139+00:00)

## What it is
LHM is a PyTorch-based large reconstruction model that reconstructs high-fidelity animatable 3D human avatars from a single image in seconds, representing them as 3D Gaussian splatting via a multimodal transformer. It includes pretrained models, a ComfyUI integration, and memory-saving variants that run on consumer GPUs.

## Use cases
- reconstruct an animatable 3D human avatar from a single photo
- generate rigged 3D digital humans for animation in seconds
- create 3D Gaussian splatting avatars without post-processing
- animate a reconstructed human with extracted motion sequences
- run 3D human reconstruction on consumer GPUs with limited memory
- integrate human reconstruction into ComfyUI workflows

## When to choose
- you need fast feed-forward animatable human reconstruction from one image
- you want 3D Gaussian splatting avatars without optimization-based refinement
- you have a consumer GPU and need memory-efficient inference
- you want a ComfyUI node for human reconstruction and animation

## When to avoid
- you need multi-view or video-based high-fidelity capture pipelines
- you require full rigging and editing in DCC tools like Blender or Maya
- you have no GPU available
- you need general object reconstruction rather than humans

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning, image-processing, graphics, simulation
- domain: computer-vision, graphics, artificial-intelligence, deep-learning
- platform: python, windows
- tags: 3d-human-reconstruction, gaussian-splatting, animatable-avatars, single-image-to-3d, digital-human, comfyui, pytorch, iccv2025, gpu, linux, docker

## Member repositories
- aigc3d/LHM (main) score 52

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:07:08.968139+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 2026-08-30T02:18:18.362544+00:00, confidence not recorded.
  - readme: https://github.com/aigc3d/LHM (fetched 2026-08-28T04:07:08.968139+00:00, sha 197efea0e3ec)
  - homepage: https://aigc3d.github.io/projects/LHM/ (fetched 2026-08-29T10:00:59.914945+00:00, sha 032e26d52789)
- Data as of 2026-08-30T08:39:29.467469+00:00.