Add prepared dataset catalog, provenance and RGB/mask/GLB previews

.gitattributes

@@ -58,3 +58,6 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 # Video files - compressed
 *.mp4 filter=lfs diff=lfs merge=lfs -text
 *.webm filter=lfs diff=lfs merge=lfs -text
+previews/3dfront/test-geometry.glb filter=lfs diff=lfs merge=lfs -text
+previews/3dfront/train-geometry.glb filter=lfs diff=lfs merge=lfs -text
+previews/ycbv/test-geometry.glb filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,119 @@
+---
+language:
+- en
+tags:
+- 3d
+- depth
+- segmentation
+- voxel-occupancy
+pretty_name: PartProbe3D preprocessed datasets
+configs:
+- config_name: 3dfront
+  data_files:
+  - split: test
+    path: catalog/3dfront/test.jsonl
+  - split: train
+    path: catalog/3dfront/train.jsonl
+- config_name: ycbv
+  data_files:
+  - split: test
+    path: catalog/ycbv/test.jsonl
+---
+
+# PartProbe3D preprocessed datasets
+
+Saved 64³ training and evaluation inputs for PartProbe3D. These archives preserve
+RGB, visible object masks, depth and validity, camera calibration, posed meshes,
+scene occupancy and independent per-object voxel occupancy. Targets describe mesh
+triangle surfaces, including hidden surfaces and overlapping object channels.
+
+| Dataset | Split | Views | Scenes |
+| --- | --- | ---: | ---: |
+| 3dfront | test | 10,066 | 935 |
+| 3dfront | train | 39,930 | 3,703 |
+| ycbv | test | 900 | 900 |
+
+3D-FRONT uses the pinned last 1,000 MIDI rooms for testing and all other rooms
+for training; all views of a room stay together. There is no validation split.
+YCB-V is evaluation only. The viewer contains metadata indices; binary training
+artifacts are in the archives.
+
+## Upload status
+
+| Dataset | Packaging complete | Upload complete | Archive GiB |
+| --- | --- | --- | ---: |
+| 3dfront | False | False | - |
+| ycbv | True | False | 3.60 |
+
+Archive inventories and SHA-256 checksums are in `bundles/<dataset>/index.json`.
+SAGE capture generation is still running and has no complete main extent dataset.
+FoundationPose currently has raw archives only. Diagnostic subsets, raw downloads,
+model weights and model-specific feature caches are not included.
+
+## Restore on another server
+
+Use the PartProbe3D checkout containing `preprocess/huggingface.py` and install its
+preprocessing dependencies plus `huggingface-hub`. Run from the repository root:
+
+```bash
+python -m preprocess.huggingface download --repo cubec/PartProbe3D-preprocessed --dataset 3dfront
+python -m preprocess.huggingface download --repo cubec/PartProbe3D-preprocessed --dataset ycbv
+```
+
+Downloads resume through Hugging Face. Restoration verifies every archive SHA-256
+and refuses to replace different existing data. `--revision <commit>` pins a release.
+The extracted roots are `datasets/3dfront/extent/point_cloud` and
+`datasets/ycbv/extent/point_cloud`. Archives are cached under each dataset’s
+`archives/huggingface/` directory. Allow space for both archives and extraction.
+
+The loaders resolve `sample`, camera, geometry and target paths relative to the
+dataset root. Historical absolute `file` and raw asset paths remain provenance;
+loading saved samples and geometry does not require those paths or raw assets.
+
+Prepare the model-specific features on the destination server, then train:
+
+```bash
+python -m ss_multiobj_depth.prepare --config ss_multiobj_depth/configs/front3d.json
+python -m ss_multiobj_depth.train --config ss_multiobj_depth/configs/front3d.json
+```
+
+## Provenance
+
+Original manifests, artifact completion markers, checksums, identities and split
+policies are retained byte-for-byte inside the archives. Generation settings and
+code hashes are also available under `metadata/<dataset>/conversion_policy.json`.
+No model checkpoint was used for this upload; no checkpoint weights are included.
+Source datasets retain their original licenses and access terms.
+
+## Input and geometry previews
+
+### 3dfront: test
+
+View: `0001acce61c945bf4470`.
+
+![RGB](previews/3dfront/test-rgb.png)
+
+![Prompt masks](previews/3dfront/test-overlay.png)
+
+[Posed prompted GT geometry (GLB)](previews/3dfront/test-geometry.glb)
+
+### 3dfront: train
+
+View: `0000b002995b4131d534`.
+
+![RGB](previews/3dfront/train-rgb.png)
+
+![Prompt masks](previews/3dfront/train-overlay.png)
+
+[Posed prompted GT geometry (GLB)](previews/3dfront/train-geometry.glb)
+
+### ycbv: test
+
+View: `003c09039641a2b2efa1`.
+
+![RGB](previews/ycbv/test-rgb.png)
+
+![Prompt masks](previews/ycbv/test-overlay.png)
+
+[Posed prompted GT geometry (GLB)](previews/ycbv/test-geometry.glb)
+

catalog.json

@@ -0,0 +1,40 @@
+{
+  "schema": "partprobe3d.hub-archives.v1",
+  "repo": "cubec/PartProbe3D-preprocessed",
+  "datasets": [
+    {
+      "name": "3dfront",
+      "source_complete": true,
+      "views": 49996,
+      "counts": {
+        "test": {
+          "views": 10066,
+          "scenes": 935
+        },
+        "train": {
+          "views": 39930,
+          "scenes": 3703
+        }
+      },
+      "source_manifest_sha256": "93fd87c7bc4e59a08b401fe83c36c02591441406f5b0535fd9e2f135b5952446",
+      "packaged": false,
+      "bytes": null,
+      "uploaded": false
+    },
+    {
+      "name": "ycbv",
+      "source_complete": true,
+      "views": 900,
+      "counts": {
+        "test": {
+          "views": 900,
+          "scenes": 900
+        }
+      },
+      "source_manifest_sha256": "5f9d0e2c94bd0d400a2852ae48e2b4656a7c939b5f8ef2816c17bea1da6a2d46",
+      "packaged": true,
+      "bytes": 3862999040,
+      "uploaded": false
+    }
+  ]
+}

metadata/3dfront/conversion_policy.json

@@ -0,0 +1,40 @@
+{
+  "schema": "partprobe3d.views",
+  "capture_manifest_sha256": "645450d21b0b894752e1eaddc9d4e0ee9f380cda7e44da93cd058bdb2ca775cc",
+  "settings": {
+    "resolution": 64,
+    "padding": 1.0,
+    "center": null,
+    "cube_size": null,
+    "minimum_mask_pixels": 64,
+    "minimum_objects": 2,
+    "max_voxels": 0,
+    "extent_strategy": "point_cloud",
+    "extent_spec": {
+      "strategy": "point_cloud",
+      "options": {
+        "depth_quantile": 0.0525,
+        "far_margin": 0.125,
+        "padding": 1.31
+      },
+      "parameters": {},
+      "mode": "train"
+    }
+  },
+  "generator_sha256": "53ceccaa63e9731641499f51ba28da74ad009232d7a4d2d0dad8ccbc8a4d4bfe",
+  "adapter_sha256": "16e5f89770c9a16ee485160e8bb710e7f6d05f8aa74666f1d3deb8899345f061",
+  "exporter_sha256": "3d84f7e02ee746e7c010086e380e120c070ca69e4442284236aae4967c54e770",
+  "camera_sha256": "3670c55189445fe07e0fec3eb57981655e9359794e186a9244fbe4251db4c674",
+  "storage_sha256": "bfef5e8d43b6f7aed38e78b4a924521c3c77ee675150c8dba1a213e5af20fdd0",
+  "dataset_sha256": "da1559a5e6ef293cd79c56142b08edaabd61806ffccc260191c0f2a2aedf6085",
+  "voxelizer_sha256": "2822855eea00a6691f6f1c4866be6267b47585cbadbc3aefe404de04306f6a84",
+  "extent_cube_sha256": "c9f7b0e24940c1c65f9398750a55e4ac730e93b784985a5583770cf94db9679f",
+  "extent_base_sha256": "3202d6f613b9766593b9a215d86e8ccd678f253d2e7c815a6bb911c2bdc253fb",
+  "ss_depth_extent_sha256": "1a7bf1d6fb62a052507706af793df3206bcb28218fbbdb6d71ee2c9a3c3ddfd6",
+  "pixal3d_extent_sha256": "e8d8d5e77511e2847f23df3c7b9ba3d7da0b18af0481378765320ff726960c9f",
+  "point_cloud_extent_sha256": "ae176a88e3b4eeb5e796da0768eb4c51f0f4578bbdcad75252670fb2c2915718",
+  "extent_factory_sha256": "e18dc4d8f9f1ef671bc537cae51877f2242bbe0a3bf513ba6a1e5f41dd186c09",
+  "extent_specification_sha256": "b0e8f0a7809192dbbfcb9da58108a632448ece4266630abafbd4571625dd4fa5",
+  "surface_voxelizer_sha256": "2cbdee866e50ad3364c2e8b62666fe33f700ca2bd2e518d10504f115b0d53f37",
+  "surface_alignment_sha256": "fa117c0f8fac4f280f4edc8e1cf8a96795e441ef5460197ffdb1840a0acaaf0d"
+}

metadata/ycbv/DATASET_CARD.md

@@ -0,0 +1,19 @@
+# YCB-V — converted per-view dataset
+
+Schema: `partprobe3d.views`. Validated **900 captures**, **900 scenes**, and **4123 binary object masks**.
+
+| Split | Scenes | Captures |
+|---|---:|---:|
+| test | 900 | 900 |
+
+Support policy: Complete CAD surfaces for the selected BOP target instances.
+
+Each `views/<id>/` contains RGB, binary masks, camera calibration, object mapping, and a uniquely named NPZ training bundle with depth and sparse voxel targets. Dense object occupancy is saved under `scenes/<id>/`; each view's `scene_gt` entry identifies its grid. Mesh-fitted bounds share one grid per scene; camera extents save one grid per view.
+
+Voxel coordinates follow the saved grid axes; model_to_world maps normalized voxel centers to the original world. Camera cubes clip surfaces outside their extent. Targets retain hidden surfaces and overlaps within the extent; solid interiors are not filled. PNG labels are one-based with zero for context. NPZ labels are zero-based with -1 for context. Depth is axial camera Z in original world units, with invalid pixels set to zero.
+
+[Full dataset index](manifest.json) · [Validation report](validation.json)
+
+Generate Markdown and GLB previews separately with `python -m preprocess.preview_dataset` under `outputs/dataset_previews/`.
+
+Use `dataloader.dataset.ViewDataset` to load arrays using portable relative paths. Full schema and converter commands are in the project's `preprocess/TWO_STAGE_DATASET.md`.

metadata/ycbv/conversion_policy.json

@@ -0,0 +1,40 @@
+{
+  "schema": "partprobe3d.views",
+  "capture_manifest_sha256": "a409d990b341a6dc8fa0a9f49a8bc557e4cf19f12e7e6107f5fe65eb806d61e6",
+  "settings": {
+    "resolution": 64,
+    "padding": 1.0,
+    "center": null,
+    "cube_size": null,
+    "minimum_mask_pixels": 64,
+    "minimum_objects": 2,
+    "max_voxels": 0,
+    "extent_strategy": "point_cloud",
+    "extent_spec": {
+      "strategy": "point_cloud",
+      "options": {
+        "depth_quantile": 0.0525,
+        "far_margin": 0.125,
+        "padding": 1.31
+      },
+      "parameters": {},
+      "mode": "train"
+    }
+  },
+  "generator_sha256": "53ceccaa63e9731641499f51ba28da74ad009232d7a4d2d0dad8ccbc8a4d4bfe",
+  "adapter_sha256": "eb81dac98f3083ebaee20c5e80561eb09d0ecfac4ffda5f4d1029931e9add08e",
+  "exporter_sha256": "3d84f7e02ee746e7c010086e380e120c070ca69e4442284236aae4967c54e770",
+  "camera_sha256": "3670c55189445fe07e0fec3eb57981655e9359794e186a9244fbe4251db4c674",
+  "storage_sha256": "bfef5e8d43b6f7aed38e78b4a924521c3c77ee675150c8dba1a213e5af20fdd0",
+  "dataset_sha256": "da1559a5e6ef293cd79c56142b08edaabd61806ffccc260191c0f2a2aedf6085",
+  "voxelizer_sha256": "2822855eea00a6691f6f1c4866be6267b47585cbadbc3aefe404de04306f6a84",
+  "extent_cube_sha256": "c9f7b0e24940c1c65f9398750a55e4ac730e93b784985a5583770cf94db9679f",
+  "extent_base_sha256": "3202d6f613b9766593b9a215d86e8ccd678f253d2e7c815a6bb911c2bdc253fb",
+  "ss_depth_extent_sha256": "1a7bf1d6fb62a052507706af793df3206bcb28218fbbdb6d71ee2c9a3c3ddfd6",
+  "pixal3d_extent_sha256": "e8d8d5e77511e2847f23df3c7b9ba3d7da0b18af0481378765320ff726960c9f",
+  "point_cloud_extent_sha256": "ae176a88e3b4eeb5e796da0768eb4c51f0f4578bbdcad75252670fb2c2915718",
+  "extent_factory_sha256": "e18dc4d8f9f1ef671bc537cae51877f2242bbe0a3bf513ba6a1e5f41dd186c09",
+  "extent_specification_sha256": "b0e8f0a7809192dbbfcb9da58108a632448ece4266630abafbd4571625dd4fa5",
+  "surface_voxelizer_sha256": "2cbdee866e50ad3364c2e8b62666fe33f700ca2bd2e518d10504f115b0d53f37",
+  "surface_alignment_sha256": "fa117c0f8fac4f280f4edc8e1cf8a96795e441ef5460197ffdb1840a0acaaf0d"
+}

metadata/ycbv/validation.json

@@ -0,0 +1,27 @@
+{
+  "complete": true,
+  "schema": "partprobe3d.views",
+  "dataset": "YCB-V",
+  "manifest_sha256": "5f9d0e2c94bd0d400a2852ae48e2b4656a7c939b5f8ef2816c17bea1da6a2d46",
+  "validated_views": 900,
+  "validated_scenes": 900,
+  "validated_grids": 900,
+  "binary_masks": 4123,
+  "counts": {
+    "test": {
+      "views": 900,
+      "scenes": 900
+    }
+  },
+  "overlapping_voxel_rows_across_views": 79629,
+  "unprompted_only_voxel_rows_across_views": 0,
+  "checks": [
+    "artifact hashes",
+    "RGB/depth/label alignment",
+    "binary mask/object identity",
+    "sparse/dense occupancy equality",
+    "shared scene GT and prompt columns",
+    "scene split separation",
+    "unique cache filenames"
+  ]
+}

previews/3dfront/test-geometry.glb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f01da3587cd40e30f1c3557c576a214cce0c19e980fad4d670f9dba91a9633c0
+size 1426740

previews/3dfront/train-geometry.glb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:946e1ea8116fdc3a89a1d50f02f565b9481a8312948509344c907c69fc707325
+size 3946796

previews/ycbv/test-geometry.glb

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a6af6426b2f3293cda10bed7c191e2b9ceeda8096e4134bf76ef3d5dbdcec8f7
+size 700796