# nv-tlabs/LLaMA-Mesh

Unifying 3D Mesh Generation with Language Models

Repository: https://github.com/nv-tlabs/LLaMA-Mesh
Canonical: https://ross.abutalabs.com/products/llama-mesh
Homepage: https://research.nvidia.com/labs/toronto-ai/LLaMA-Mesh/
Language: Python
License: NOASSERTION
License Family: other
Topics: 3d-generation, llm, mesh-generation, multimodal
Last push: 2025-03-28T07:00:47+00:00

## Health v2 (maintenance only)
Score: 28/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 13, release rhythm 35, longevity 47
- inputs: {"age_days": 658, "days_push": 523, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1166, forks 77 (observed 2026-08-28T04:03:50.204637+00:00)

## What it is
LLaMA-Mesh is a fine-tuned large language model from NVIDIA Research that generates and understands 3D meshes by representing vertex coordinates and face definitions as plain text. It supports conversational text-to-3D generation, interleaved text and mesh output, and mesh interpretation, with a Gradio demo UI and a Blender addon.

## Use cases
- generate 3d mesh from a text prompt
- chat with an llm to create 3d models
- convert text descriptions into obj mesh files
- understand and interpret 3d meshes with a language model
- fine-tune an llm for 3d asset generation
- generate 3d models inside blender via addon

## When to choose
- you want conversational, prompt-driven 3D mesh generation in a single model
- you need a model that keeps strong text abilities while outputting 3D meshes
- you want to experiment with text-based 3D tokenization and LLM fine-tuning research

## When to avoid
- you need high-fidelity production 3D assets with detailed textures
- you lack a GPU or cannot run large language model inference locally
- you need non-mesh 3D formats or CAD-precision geometry

## Facets
- artifact type: library
- maturity: active
- function: llm-inference, machine-learning, deep-learning, llm-training
- domain: large-language-models, artificial-intelligence, graphics
- platform: python
- tags: 3d-generation, mesh-generation, multimodal, text-to-3d, obj-format, gradio, nvidia, fine-tuning, transformers, game-development, gpu, linux

## Member repositories
- nv-tlabs/LLaMA-Mesh (main) score 28

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:03:50.204637+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 2026-08-30T06:29:17.598030+00:00, confidence not recorded.
  - readme: https://github.com/nv-tlabs/LLaMA-Mesh (fetched 2026-08-28T04:03:50.204637+00:00, sha d9f7bb41870a)
  - homepage: https://research.nvidia.com/labs/toronto-ai/LLaMA-Mesh/ (fetched 2026-08-29T12:35:13.561782+00:00, sha 7d968bc7971e)
- Data as of 2026-08-30T08:39:29.467469+00:00.