# thu-ml/prolificdreamer

ProlificDreamer: High-Fidelity and Diverse Text-to-3D Generation with Variational Score Distillation (NeurIPS 2023 Spotlight)

Repository: https://github.com/thu-ml/prolificdreamer
Canonical: https://ross.abutalabs.com/products/prolificdreamer
Homepage: https://ml.cs.tsinghua.edu.cn/prolificdreamer/
Language: Python
License: Apache-2.0
License Family: permissive
Topics: diffusion-model, dreamfusion, nerf, text-to-3d, prolificdreamer, stablediffusion
Last push: 2023-11-22T04:42:56+00:00

## Health v2 (maintenance only)
Score: 29/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 35, longevity 85
- inputs: {"age_days": 1196, "days_push": 1015, "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 1564, forks 43 (observed 2026-08-28T04:05:04.141176+00:00)

## What it is
Official PyTorch implementation of ProlificDreamer, a NeurIPS 2023 method for high-fidelity text-to-3D generation using Variational Score Distillation (VSD) with pretrained Stable Diffusion. It generates NeRF scenes in a multi-stage pipeline and refines them into detailed, textured meshes via DMTet.

## Use cases
- generate 3d models from text prompts
- create textured meshes from a text description
- train a nerf from text using stable diffusion guidance
- research variational score distillation for text-to-3d
- compare sds vs vsd for 3d generation
- produce high-resolution 3d assets for games or rendering

## When to choose
- you need state-of-the-art text-to-3d quality with detailed meshes
- you are researching score distillation methods like VSD vs SDS
- you have a high-memory GPU (20-27GB+) and want to reproduce paper results

## When to avoid
- you need fast or real-time 3d generation (training takes many GPU-hours)
- you lack a large-VRAM GPU
- you need production-ready tooling rather than research code
- you cannot tolerate the multi-face Janus artifact common to SDS/VSD methods

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning, graphics, simulation
- domain: artificial-intelligence, computer-vision, graphics, deep-learning
- platform: python
- tags: text-to-3d, nerf, diffusion-models, stable-diffusion, vsd, score-distillation, dmtet, research-code, neurips-2023, linux, gpu

## Member repositories
- thu-ml/prolificdreamer (main) score 29

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:04.141176+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 2026-08-30T04:00:14.213506+00:00, confidence not recorded.
  - readme: https://github.com/thu-ml/prolificdreamer (fetched 2026-08-28T04:05:04.141176+00:00, sha 3cdab4222a39)
  - homepage: https://ml.cs.tsinghua.edu.cn/prolificdreamer/ (fetched 2026-08-29T11:29:07.554381+00:00, sha 0c7365a3c9eb)
- Data as of 2026-08-30T08:39:29.467469+00:00.