This dataset contains the generated artifacts for the first 40 MeshEval/Tripo cases evaluated with PointLLM-R.
mesheval_tripo_first40/manifest.json: case metadata and point-cloud provenance.
mesheval_tripo_first40/points/: sampled 8192-point XYZRGB NumPy arrays.
mesheval_tripo_first40/results/: one structured geometry evaluation JSON per case.
mesheval_tripo_first40/report/: self-contained static dashboard… See the full description on the dataset page:
https://huggingface.co/datasets/blllank/PointLLM-R-eval.