Upload README.md with huggingface_hub

README.md

@@ -39,11 +39,12 @@ from affostruction import AffostructionPipeline
 pipeline = AffostructionPipeline.from_pretrained("chrockey/Affostruction").cuda()
 outputs = pipeline.run(input_dict, queries=["grasp"])
 
-mesh   = outputs["mesh"][0]
 coords = outputs["coords"]                  # (N, 4) sparse voxel coords
 probs  = outputs["affordance"][0]["probs"]  # (N,) per-voxel heatmap in [0, 1] (paired with coords)
 ```
 
+> **Want a mesh / gaussian?** Pass `formats=["mesh", "gaussian"]` to `pipeline.run` to decode renderable outputs. Decoding is opt-in to avoid pulling in nvdiffrast / kaolin when callers only need the heatmap.
+>
 > **Reconstruction only:** drop `queries` to skip affordance, or pass `load_affordance=False` to `from_pretrained` to avoid downloading the affordance denoiser.
 
 See `examples/affordance.py` (full pipeline) and `examples/reconstruction.py` / `examples/reconstruction_unposed.py` (reconstruction-only) in the [GitHub repo](https://github.com/chrockey/Affostruction).