Add AI2-Thor Perspective Taking QA Dataset with dual image capture and enhanced metadata

README.md

@@ -0,0 +1,149 @@
+---
+dataset_info:
+  features:
+  - name: question
+    dtype: string
+  - name: answer
+    dtype: string
+  - name: choices
+    sequence: string
+  - name: marked_image
+    dtype: image
+  - name: new_perspective
+    dtype: image
+  - name: scene_name
+    dtype: string
+  - name: trajectory_id
+    dtype: string
+  - name: question_type
+    dtype: string
+  - name: difficulty
+    dtype: string
+  - name: movement_type
+    dtype: string
+  - name: platform
+    dtype: string
+  - name: metadata
+    struct:
+    - name: generation_method
+      dtype: string
+    - name: has_marked_image
+      dtype: bool
+    - name: has_new_perspective
+      dtype: bool
+    - name: original_question_with_choices
+      dtype: string
+    - name: question_metadata
+      struct:
+      - name: agent_states
+        struct:
+        - name: initial
+          struct:
+          - name: cameraHorizon
+            dtype: float64
+          - name: position
+            struct:
+            - name: x
+              dtype: float64
+            - name: y
+              dtype: float64
+            - name: z
+              dtype: float64
+          - name: rotation
+            struct:
+            - name: x
+              dtype: float64
+            - name: y
+              dtype: float64
+            - name: z
+              dtype: float64
+        - name: new
+          struct:
+          - name: cameraHorizon
+            dtype: float64
+          - name: position
+            struct:
+            - name: x
+              dtype: float64
+            - name: y
+              dtype: float64
+            - name: z
+              dtype: float64
+          - name: rotation
+            struct:
+            - name: x
+              dtype: float64
+            - name: y
+              dtype: float64
+            - name: z
+              dtype: float64
+      - name: answer_choices
+        sequence: string
+      - name: image_paths
+        struct:
+        - name: new_perspective
+          dtype: string
+        - name: original_marked
+          dtype: string
+      - name: movement_info
+        struct:
+        - name: direction
+          dtype: string
+        - name: rotation_change
+          dtype: float64
+        - name: target_point_2d
+          sequence: float64
+        - name: target_position_3d
+          struct:
+          - name: x
+            dtype: float64
+          - name: y
+            dtype: float64
+          - name: z
+            dtype: float64
+      - name: object_info
+        struct:
+        - name: distance_change
+          dtype: float64
+        - name: initial_distance
+          dtype: float64
+        - name: new_distance
+          dtype: float64
+        - name: object_id
+          dtype: string
+      - name: original_compatibility
+        dtype: string
+      - name: original_question
+        dtype: string
+      - name: question_index
+        dtype: int64
+      - name: sample_index
+        dtype: int64
+      - name: scene_name
+        dtype: string
+      - name: template
+        dtype: string
+    - name: task_type
+      dtype: string
+  splits:
+  - name: train
+    num_bytes: 2609201585.8
+    num_examples: 1620
+  - name: validation
+    num_bytes: 330083637.59358025
+    num_examples: 202
+  - name: test
+    num_bytes: 340617877.60641974
+    num_examples: 203
+  download_size: 2943477577
+  dataset_size: 3279903101.0
+configs:
+- config_name: default
+  data_files:
+  - split: train
+    path: data/train-*
+  - split: validation
+    path: data/validation-*
+  - split: test
+    path: data/test-*
+---

data/test-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e688526b06b3bf1477d08dc191fb32166cdffcf5d1bd1cc8713f515677c30c34
+size 287836191

data/train-00000-of-00006.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:30f0df692111154b5ce7653d2d22e2b79c275f5bd402a52e621f83bae5c3e900
+size 411107038

data/train-00001-of-00006.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b910713a438d2bea4bdf5b30557eff80e9c354eaa598f52dfbf920e974b80839
+size 394595903

data/train-00002-of-00006.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6ed923c68559f50802f9c19ee6799a2bd73e9921cd7124324e4e8241113a9e8c
+size 384963248

data/train-00003-of-00006.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aa3e0478da0a4575403e873a3392239933ecd2b9253b7121523de84808adc069
+size 389935297

data/train-00004-of-00006.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e6651b572ffa9233badee81d78cf5707449df718955522d5fbfcad62ca76c86e
+size 386973195

data/train-00005-of-00006.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3e8044e979bb2343b5add7563a3d6e2e2b328060119ab6acf72f6b61401bac18
+size 396403631

data/validation-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e0eadf026c971c4c3530cd29202fc3caf99224c35f90485a70c8449905dab1e6
+size 291663074