The shap-e repository provides the official code and model release for Shap-E, a conditional generative model designed to produce 3D assets (implicit functions, meshes, neural radiance fields) from text or image prompts. The model is built with a two-stage architecture: first an encoder that maps existing 3D assets into parameterizations of implicit functions, and then a conditional diffusion model trained on those parameterizations to generate new assets. Because it works at the level of implicit functions, Shap-E can render output both as textured meshes and NeRF-style volumetric renderings. The repository contains sample notebooks (e.g. sample_text_to_3d.ipynb, sample_image_to_3d.ipynb) so users can try out text → 3D or image → 3D generation. The code is distributed under the MIT license, and includes a “model card” that documents limitations, recommended use, and ethical considerations.
Features
- Conditional generation of 3D implicit function models from text or images
- Two-stage model architecture: encoder + diffusion over implicit parameter space
- Output in multiple representations: meshes, NeRF renderings
- Sample notebooks for text23D and image23D use cases
- Model card documenting limitations, biases, and usage guidance
- MIT-licensed code, allowing reuse and extension
Project Samples
