microsoft/trellis.2 ? reverse-engineered prompt

Build me a simple app for TRELLIS.2 that lets me upload an image and turn it into a high quality 3D model with textures.

I want a clean interface where I can drop in a picture, generate the 3D result, preview it in the browser, and then download it as a GLB file. If it makes sense, also add a second mode for generating or improving textures from an existing shape. Please keep the experience easy to use, with a clear button to start generation, a progress indicator, and a place to view the rendered output before saving it.

Use the pretrained model that comes with this project, and make sure it works smoothly on a GPU machine. If you need to look up the current docs or best setup for the model, go ahead and do that. I’d also like the app to be wired up so I can run it locally without a lot of extra steps.