>>106175506
Don't see anywhere in the code that allows for multi view easily, but it uses similar diffusion steps to other image-3d generator projects. It could probably be done since part of its inference steps is to generate multiple views of the image before running the mesh generation but I'm not bothered to dig into the code to add it since it's been able to handle it very well with just 1 image reference.
I've ran trellis and it had the ability to specify multiple views but it has issues with baking lighting into the textures. I prefer Hunyuan3D 2.1 so far.