Emma

Summary:

  • Point-E is a new method for generating 3D models from text prompts, producing results in 1-2 minutes on a single GPU.
  • Though sample quality is lower than state-of-the-art methods, its speed offers a practical trade-off for certain use cases.