Name the unseen-reference tracker explicitly

README.md

@@ -29,7 +29,7 @@ Dexterous Manipulation Tracking**, CVPR 2026.
 |---|---|---|
 | `assets.tar.gz` | `assets/` | Object meshes and URDFs, xArm/LEAP robot model, floor texture, object latents |
 | `references.tar.gz` | `data/adadextrack_full/` | All 2,765 reference motions, manifests, semantic groups, split lists and provenance |
-| `bc_tracker.pth` | `checkpoints/bc_tracker.pth` | Pretrained unseen-trajectory BC tracker, 273 observations and 22 actions |
+| `bc_tracker_unseen_ref.pth` | `checkpoints/bc_tracker_unseen_ref.pth` | Pretrained unseen-trajectory BC tracker, 273 observations and 22 actions |
 | `bc_tracker_unseen_obj.pth` | `checkpoints/bc_tracker_unseen_obj.pth` | Pretrained unseen-object BC tracker, 273 observations and 22 actions |
 | `release.json` | Download metadata | Archive and per-file SHA256 checksums, sizes and coverage |
 
@@ -49,7 +49,7 @@ transforms. Position units are meters, angles are radians and quaternions are xy
 The original NumPy dictionaries use pickle serialization; the project loader
 reads these project resource files with `allow_pickle=True`.
 
-Two BC trackers are released: `bc_tracker.pth` is trained on the 2,212-reference
+Two BC trackers are released: `bc_tracker_unseen_ref.pth` is trained on the 2,212-reference
 unseen-trajectory training split and is evaluated on `unseen_trajectory_test.csv`
 (553 references); `bc_tracker_unseen_obj.pth` is trained on the 2,520-reference
 unseen-object training split and is evaluated on `unseen_object_test.csv`
@@ -78,7 +78,7 @@ python scripts/download_resources.py --components assets
 ```
 
 For manual installation, extract both `.tar.gz` archives into the **code
-repository root** and put both `bc_tracker.pth` and
+repository root** and put both `bc_tracker_unseen_ref.pth` and
 `bc_tracker_unseen_obj.pth` in `checkpoints/`. Archives already
 include their top-level `assets/` and `data/adadextrack_full/` paths.
 Isaac Gym Preview 4 must be obtained and installed separately.
@@ -102,7 +102,7 @@ Run each tracker with its matching test split from the code repository root:
 
 ```bash
 python scripts/infer_bc.py \
-  --checkpoint checkpoints/bc_tracker.pth \
+  --checkpoint checkpoints/bc_tracker_unseen_ref.pth \
   --ref-csv data/adadextrack_full/splits/unseen_trajectory_test.csv \
   --batch-size 553 --device cuda:0 --seed 42
 
@@ -113,7 +113,7 @@ python scripts/infer_bc.py \
 ```
 
 Add `--render` for a viewer. `python scripts/download_resources.py` downloads
-both checkpoints by default; use `--components bc_tracker bc_tracker_unseen_obj`
+both checkpoints by default; use `--components bc_tracker_unseen_ref bc_tracker_unseen_obj`
 to download only the two weights.
 
 The code uses L2 position errors, quaternion angular errors and 16-finger-joint

release.json

@@ -2,7 +2,7 @@
   "schema_version": 1,
   "repo_id": "Janebek/AdaDexTrack",
   "repo_type": "dataset",
-  "created_utc": "2026-10-23T14:39:02.874252+00:00",
+  "created_utc": "2026-10-23T14:58:57.096626+00:00",
   "dataset": {
     "name": "AdaDexTrack-Full",
     "references": 2765,
@@ -13953,13 +13953,13 @@
         }
       }
     },
-    "bc_tracker": {
+    "bc_tracker_unseen_ref": {
       "bytes": 196865446,
       "sha256": "b4e918e91c1c974b945725b469354662b25924abf4d744e3805299864a980c33",
-      "filename": "bc_tracker.pth",
+      "filename": "bc_tracker_unseen_ref.pth",
       "format": "file",
       "files": {
-        "checkpoints/bc_tracker.pth": {
+        "checkpoints/bc_tracker_unseen_ref.pth": {
           "bytes": 196865446,
           "sha256": "b4e918e91c1c974b945725b469354662b25924abf4d744e3805299864a980c33"
         }
@@ -13980,8 +13980,8 @@
   },
   "metadata_files": {
     "README.md": {
-      "bytes": 5636,
-      "sha256": "924f52048489c8d33689b3f90e2b9dbe66767ffccc8b315354f1e7ebfe9642f1"
+      "bytes": 5702,
+      "sha256": "c2686dfbba3c27bb40bf62a5c45bed3dbe607eba95c6a10922151a4ffea6f5e8"
     },
     "RESOURCE_LICENSES.md": {
       "bytes": 2969,