Open-AI recently released Shap-E, which can create object meshes from both text & image (or per their words: “conditional generative model for 3D assets”).

• Github Repo: https://github.com/openai/shap-e
• Paper describing Shap-E from OpenAI

On this page…

• Setting up Shap-E
• Text to 3D
• Image to 3D
• Learnings
• References

<aside> 💡 This experiment was done on 11 May, 2023. It took me about 90 minutes to do all the below from scratch, and write this post along the way!

</aside>

Setting up Shap-E

Shap-E

Learnings

References