Stable Dreamfusion

Generate a 3D object (*.obj, *.mtl, albedo.png) from a single text input, based on the Stable Diffusion Model. This project is a work-in-progress, and contains lots of differences from the paper Dreamfusion↗︎. Also, many features are still not implemented now. The current generation quality cannot match the results from the original paper, and many prompts still fail badly! Generation takes about 1.5 hrs on T4 on Google Colab.