Split target feature

README.md

@@ -5,32 +5,32 @@ dataset_info:
   features:
   - name: img
     dtype: image
-  - name: target
+  - name: labels
+    sequence:
+      class_label:
+        names:
+          '0': Dog
+          '1': Motorbike
+          '2': People
+          '3': Cat
+          '4': Chair
+          '5': Table
+          '6': Car
+          '7': Bicycle
+          '8': Bottle
+          '9': Bus
+          '10': Cup
+          '11': Boat
+  - name: bboxes
     sequence:
-    - name: labels
-      dtype:
-        class_label:
-          names:
-            '0': Dog
-            '1': Motorbike
-            '2': People
-            '3': Cat
-            '4': Chair
-            '5': Table
-            '6': Car
-            '7': Bicycle
-            '8': Bottle
-            '9': Bus
-            '10': Cup
-            '11': Boat
-    - name: bboxes
       sequence: float32
+      length: 4
   splits:
   - name: train
-    num_bytes: 1864877
+    num_bytes: 1770065
     num_examples: 7361
   download_size: 1487935964
-  dataset_size: 1864877
+  dataset_size: 1770065
 ---
 
 The Exclusively Dark (ExDARK) dataset is a collection of low-light

ex-dark.py

@@ -1,4 +1,5 @@
-# Copyright 2020 The HuggingFace Datasets Authors and the current dataset script contributor.
+# Copyright 2020 The HuggingFace Datasets Authors and the current dataset
+# script contributor.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -103,16 +104,17 @@ class ExDark(datasets.GeneratorBasedBuilder):
             features=datasets.Features(
                 {
                     "img": datasets.Image(),
-                    "target": datasets.Sequence(
-                        feature={
-                            "labels": datasets.features.ClassLabel(
-                                names=_LABEL_NAMES
-                            ),
-                            "bboxes": datasets.features.Sequence(
-                                feature=datasets.Value("float32")
-                            )
-                        }
-                    )
+                    "labels": datasets.Sequence(
+                        feature=datasets.features.ClassLabel(
+                            names=_LABEL_NAMES,
+                        ),
+                    ),
+                    "bboxes": datasets.Sequence(
+                        feature=datasets.Sequence(
+                            feature=datasets.Value("float32"),
+                            length=4,
+                        ),
+                    ),
                 }
             ),
             homepage=_HOMEPAGE,
@@ -149,8 +151,6 @@ class ExDark(datasets.GeneratorBasedBuilder):
             labels = sample["class"].to_numpy()
             yield idx, {
                 "img": img_path,
-                "target": {
-                    "labels": labels,
-                    "bboxes": bboxs,
-                },
+                "labels": labels,
+                "bboxes": bboxs,
             }