سيفار

  • الوصف :

نسخة معاد تسميتها من مجموعة اختبار CIFAR-10 مع ملصقات إلكترونية قادمة من مشروحين بشريين حقيقيين. لكل زوج (صورة، ملصق) في مجموعة اختبار CIFAR-10 الأصلية، فإنه يوفر العديد من التسميات الإضافية التي يقدمها المفسرون البشريون الحقيقيون بالإضافة إلى التسمية الناعمة المتوسطة. مجموعة التدريب مطابقة لمجموعة البيانات الأصلية.

ينقسم أمثلة
'test' 10.000
'train' 50.000
  • هيكل الميزة :
FeaturesDict({
   
'annotator_ids': Sequence(Scalar(shape=(), dtype=int32)),
   
'human_labels': Sequence(ClassLabel(shape=(), dtype=int64, num_classes=10)),
   
'id': Text(shape=(), dtype=string),
   
'image': Image(shape=(32, 32, 3), dtype=uint8),
   
'label': ClassLabel(shape=(), dtype=int64, num_classes=10),
   
'reaction_times': Sequence(Scalar(shape=(), dtype=float32)),
   
'soft_label': Tensor(shape=(10,), dtype=float32),
   
'trial_indices': Sequence(Scalar(shape=(), dtype=int32)),
})
  • وثائق الميزة :
ميزة فصل شكل نوع D وصف
المميزاتDict
annotator_ids تسلسل (عددي) (لا أحد،) int32
human_labels التسلسل (ClassLabel) (لا أحد،) int64
بطاقة تعريف نص خيط
صورة صورة (32، 32، 3) uint8
ملصق ClassLabel int64
Reaction_times تسلسل (عددي) (لا أحد،) float32
soft_label الموتر (10،) float32
محاكمة_مؤشرات تسلسل (عددي) (لا أحد،) int32

التصور

  • الاقتباس :
@inproceedings{wei2022learning,
  title
={Human uncertainty makes classification more robust},
  author
={Joshua C. Peterson and Ruairidh M. Battleday and Thomas L. Griffiths
 
and Olga Russakovsky},
  booktitle
={IEEE International Conference on Computer Vision and Pattern
 
Recognition (CVPR)},
  year
={2019}
}