now most of the new models are concentrating on text-to-3D or single image-to-3D generation, meaning they do generate not exact object that people want. to fix this we use as input images from the main views of the object and generate exact 3D models.