Upload README.md with huggingface_hub

README.md

@@ -0,0 +1,61 @@
+---
+license: other
+license_name: source-licenses
+license_details: Each asset remains under the license of its original distribution (Articraft, PartNeXt, Fusion360, TRELLIS-500K sources).
+task_categories:
+  - image-to-3d
+tags:
+  - 3d
+  - part-aware
+  - training-data
+---
+
+# PartTrellis corpus — 4-source part-labelled training data
+
+The full training and test corpus of PartTrellis: **19,132 training objects**
+(`lists/tri_train_dual.txt`) and **1,000 held-out test objects**
+(`lists/test1000.txt`), drawn uniformly from four part-labelled sources:
+Articraft, PartNeXt, Fusion360, and TRELLIS-500K assets. Includes the original
+textured meshes, the rendered conditioning views, and all precomputed training
+targets (O-Voxel dual volumes and latents), so both training and evaluation can
+run without re-doing preprocessing.
+
+## Layout
+
+```
+shards/assets_000.tar …          # ~1,000 objects each (train ∪ test)
+  <source>/<oid>/mesh.glb            original textured mesh
+  <source>/<oid>/face_ids.npy        per-face part labels
+  <source>/<oid>/meta.json
+  cond/<source>/<oid>/renders_cond_pbr/   PBR conditioning views + transforms.json
+
+latents/latents_000.tar …        # 5,000 training objects each
+  <source>/<oid>/A.vxz  B.vxz        packed dual O-Voxel volumes (Stage-1 target)
+  <source>/<oid>/tri_ss_latent.npz   sparse-structure latent
+  <source>/<oid>/tri_slat.npz        SLat latent (legacy fitted-VAE space)
+  <source>/<oid>/tri_meta.json       bipartite contraction metadata
+latents/origslat.tar             # SLat latents in the ORIGINAL TRELLIS.2 VAE
+                                 # space — the Stage-2 target of the released model
+
+eval/gt_cache.tar                # cached GT part point samples for the evaluator
+lists/                           # train/test/per-source curation lists + recipes
+```
+
+Note on Stage-2 latents: the released PartTrellis Stage-2
+(`stage2_slat_orig_100k.pt`) trains on `latents/origslat.tar` (original
+TRELLIS.2 VAE space). `tri_slat.npz` inside the latent shards is the legacy
+fitted-VAE space kept for reproducing older ablations; do not mix the two.
+
+## Preprocessing provenance
+
+Meshes are voxelized **as-is** after rigid normalization only (centering, unit
+cube, Z-up) — O-Voxel is a field-free surface proxy, so no watertighting,
+remeshing, or cleaning is applied. Part contact graphs are contracted until
+bipartite and packed into two volumes (`A.vxz`/`B.vxz`). All preprocessing
+scripts ship in the code tarball of
+https://huggingface.co/AuroraRyan/PartTrellis (`attraction/data/`).
+
+## Licensing
+
+The packaging and lists are released for research use; every asset remains
+under the license of its original distribution.