Datasets:

Languages:
English
Multilinguality:
monolingual
Size Categories:
10K<n<100K
Language Creators:
crowdsourced
Annotations Creators:
crowdsourced
ArXiv:
Tags:
toloka
License:
Senarect commited on
Commit
0066ca3
1 Parent(s): 7955c2e

Update read.me for better description

Browse files
Files changed (1) hide show
  1. README.md +58 -0
README.md CHANGED
@@ -70,3 +70,61 @@ Our dataset has 45,199 instances split among three subsets: train (38,990 instan
70
  and private test (4,504 instances). The entire train dataset was available for everyone since the start of the challenge.
71
  The public test dataset was available since the evaluation phase of the competition, but without any ground truth labels.
72
  After the end of the competition, public and private sets were released.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
70
  and private test (4,504 instances). The entire train dataset was available for everyone since the start of the challenge.
71
  The public test dataset was available since the evaluation phase of the competition, but without any ground truth labels.
72
  After the end of the competition, public and private sets were released.
73
+
74
+ ### Supported Tasks and Leaderboards
75
+
76
+ The Visual Question Answering.
77
+
78
+ ### Language
79
+
80
+ English
81
+
82
+ ## Dataset Structure
83
+
84
+ ### Data Instances
85
+ A data instance contains a url to the picture, information about the image size - width and height,
86
+ information about ground truth bounding box - left top and right bottom dots,
87
+ contains the question related to the picture.
88
+ image,width,height,left,top,right,bottom,question
89
+ ```
90
+ {'image': https://toloka-cdn.azureedge.net/wsdmcup2023/000000000013.jpg,
91
+ 'width': 640,
92
+ 'height': 427,
93
+ 'left': 129,
94
+ 'top': 192,
95
+ 'right': 155,
96
+ 'bottom': 212,
97
+ 'question': What does it use to breath?}
98
+ ```
99
+ ### Data Fields
100
+
101
+ * image: contains url to the image
102
+ * width: value in pixels of image width
103
+ * heigth: value in pixels of image height
104
+ * left: the x coordinate in pixels to determin left-top dot of bounding box
105
+ * top: the y coordinate in pixels to determin left-top dot of bounding box
106
+ * right: the x coordinate in pixels to determin right-bottom dot of bounding box
107
+ * bottom: the y coordinate in pixels to determin right-bottom dot of bounding box
108
+ * question: a question related to the picture
109
+
110
+ ### Data Splits
111
+ There are four splits in the data: train, train_sample, test_public, test_private. 'train' split contains the full pull for model training.
112
+ 'train-sample' split contains the part of 'train' split. 'test_public' split contains public data to test the model.
113
+ 'test_private' split contains private data for final model test.
114
+
115
+ ### Source Data
116
+
117
+ The images were obtained from a CC BY-licensed subset of the Microsoft Common Objects in
118
+ Context dataset, [MS COCO](https://cocodataset.org/).
119
+
120
+ ### Annotations
121
+
122
+ All data labeling was performed on the [Toloka crowdsourcing platform](https://toloka.ai/).
123
+
124
+ Only annotators who self-reported the knowledge of English had access to the annotation task.
125
+
126
+ ### Citation Information
127
+
128
+ * Competition: https://toloka.ai/challenges/wsdm2023
129
+ * CodaLab: https://codalab.lisn.upsaclay.fr/competitions/7434
130
+ * Dataset: https://doi.org/10.5281/zenodo.7057740