Append 2,410 vision training rows; preserve previous data and evaluation

CHANGELOG.md

@@ -4,6 +4,11 @@ Changes that affect published data, newest first: one entry per publish, `## <da
 concise bullets on what changed in the data. `tasksource-jev-typed-decisions` releases are tagged on the Hub
 (`load_dataset("tasksource/tasksource-jev-typed-decisions", revision="v2026.10.13")`).
 
+## tasksource-jev-typed-decisions v2026.10.16
+
+- `vision`: add `spatial-dise/mcq` with native spatial-reasoning choices and original composite images. Training excludes generator families reserved by validation, test or the separate Spatial-DISE benchmark.
+- Retain rotation, combination and hidden-face questions. Projection is omitted for missing viewing cues; folding for a sampled ambiguous answer set. Source terms are CC BY-NC-SA 4.0 with original third-party asset rights retained.
+
 ## tasksource-jev-typed-decisions v2026.10.15.5
 
 - `vision`: removed the grid-cell grounding rows (rico-widget/grid7, spair71k/grid7); the coordinate-option views of v2026.10.15.4 replace them. Other rows and shards are unchanged.

README.md

@@ -207,16 +207,16 @@ dataset_info:
     dtype: string
   splits:
   - name: train
-    num_bytes: 222333380352
-    num_examples: 659904
+    num_bytes: 223497244350
+    num_examples: 662314
   - name: test
     num_bytes: 146575133
     num_examples: 475
   - name: validation
     num_bytes: 274444703
     num_examples: 972
-  download_size: 132672233952
-  dataset_size: 222754400188
+  download_size: 133832193630
+  dataset_size: 223918264186
 ---
 
 # tasksource-jev-typed-decisions

vision/release-audit.json

@@ -3,7 +3,7 @@
   "config": "vision",
   "splits": {
     "train": {
-      "rows": 659904,
+      "rows": 662314,
       "sources": {
         "vision/tallyqa/count": 30000,
         "vision/nlvr2": 30000,
@@ -49,7 +49,8 @@
         "vision/rico-widget/box": 30000,
         "vision/mind2web/box": 5086,
         "vision/mind2web/element": 5086,
-        "vision/mind2web/action": 5086
+        "vision/mind2web/action": 5086,
+        "vision/spatial-dise/mcq": 2410
       },
       "families": {
         "vision/tallyqa": 30000,
@@ -80,11 +81,12 @@
         "vision/sphinx": 30000,
         "vision/trace": 30000,
         "vision/visgym": 30000,
-        "vision/mind2web": 15258
+        "vision/mind2web": 15258,
+        "vision/spatial-dise": 2410
       },
       "license_use": {
         "unspecified": 151691,
-        "non-commercial": 134206,
+        "non-commercial": 136616,
         "commercial": 374007
       }
     },
@@ -433,6 +435,10 @@
     {
       "manifest": "vision/additions/6296521bd9e3/manifest.json",
       "train_rows": 75258
+    },
+    {
+      "manifest": "vision/additions/bc2291cb6835/manifest.json",
+      "train_rows": 2410
     }
   ],
   "removals": [

vision/sources.yaml

@@ -7,6 +7,7 @@ datasets:
 - HuggingFaceM4/the_cauldron
 - OpenMOSS-Team/GameQA-140K
 - RyanWW/Super-CLEVR
+- TACPS-liv/Spatial-DISE
 - Valen-Team/Valen-Eval-Game
 - bevaya/RICO-WidgetCaptioning
 - derek-thomas/ScienceQA
@@ -26,6 +27,7 @@ datasets:
 - tasksource/mentis-oculi-paper-fold
 - tasksource/multimodal-mind2web
 - tasksource/spair71k-grid
+- tasksource/spatial-dise
 - tasksource/sphinx-mcq
 - tasksource/superclevr
 - tasksource/trace-mcq
@@ -840,4 +842,30 @@ sources:
       - openrail
       osunlp/Multimodal-Mind2Web:
       - openrail
+  vision/spatial-dise/mcq:
+    rows:
+      train: 2410
+    dataset: tasksource/spatial-dise
+    revision: 6d597407f3b15b86cf2c323373ac9bf1c5065da9
+    originals:
+    - TACPS-liv/Spatial-DISE
+    revisions:
+      tasksource/spatial-dise: 6d597407f3b15b86cf2c323373ac9bf1c5065da9
+    license: cc-by-nc-sa-4.0, cc-by-nc-sa-4.0 (original source)
+    license_use: non-commercial
+    card_licenses:
+      tasksource/spatial-dise:
+      - cc-by-nc-sa-4.0
+      TACPS-liv/Spatial-DISE:
+      - cc-by-nc-sa-4.0
+    source_license_evidence:
+      license: cc-by-nc-sa-4.0
+      license_use: non-commercial
+      checked: '2026-10-16'
+      scope: Spatial-DISE release; Icons8 textures and collected aptitude-test images
+        retain original third-party terms
+      url: https://huggingface.co/datasets/TACPS-liv/Spatial-DISE#license
+      additional_urls:
+      - https://arxiv.org/html/2510.13394v1#A3
+      - https://arxiv.org/html/2510.13394v1#A4
 parent_tasksource_commit: f9962e198e653bfb73cc2ae5ea33e1511b6cb746

vision/train-append-bc2291cb6835-00000-of-00005.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:24ecb30228c1180d87429d453b87a17bb384125f7e3b7646797a466bc45132e5
+size 232019614

vision/train-append-bc2291cb6835-00001-of-00005.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a4530268a9b45ecb0f5fad8b8c9468ef98a6bd693394a9db0d9af8a0696b044c
+size 225783518

vision/train-append-bc2291cb6835-00002-of-00005.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5a710871d1ea0b209a944b2cb8fd650afeae33ab9d11d5ac07557f5211f4eff4
+size 234964492

vision/train-append-bc2291cb6835-00003-of-00005.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e5010f2e93699445eb18e43fe431452eb4e031ca4aec94d012eb6e636d27e241
+size 230041812

vision/train-append-bc2291cb6835-00004-of-00005.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:383df151a8dab6d4427f0585455254308b504a54a5de2d5741b124bf5501891d
+size 237150242