Upload README.md with huggingface_hub

README.md

@@ -47,6 +47,6 @@ probs  = outputs["affordance"][0]["probs"]  # (N,) per-voxel heatmap in [0, 1],
 >
 > **Reconstruction only?** Drop `queries` (or pass `load_affordance=False` to `from_pretrained`).
 >
-> **Active view selection.** Pass `view_selection_mode="voxel"` (decoder-free) or `"mesh"` (paper version, needs `nvdiffrast`). Candidate poses come from a dataset `transforms.json` (`view_selection_transforms_path=...`, Blender c2w → OpenCV w2c conversion baked in).
+> **Active view selection.** On by default in `"voxel"` mode (per-pixel raycast, decoder-free). Pass `view_selection_mode="mesh"` for the paper version (SLAT mesh decode + nvdiffrast), or `None` to skip. Candidate poses come from a dataset `transforms.json` (`view_selection_transforms_path=...`, Blender c2w → OpenCV w2c conversion baked in).
 
 End-to-end examples in the [GitHub repo](https://github.com/chrockey/Affostruction): `examples/affostruction.py` (full pipeline), `examples/reconstruction.py` (reconstruction-only), and `examples/reconstruction_unposed.py` (experimental unposed single-view).