- وصف :
ImageNet-PI هي نسخة معاد تسميتها من مجموعة بيانات ILSVRC2012 ImageNet القياسية، حيث يتم توفير التصنيفات بواسطة مجموعة من 16 شبكة عصبية عميقة ذات بنى مختلفة تم تدريبها مسبقًا على مجموعة بيانات ILSVRC2012 القياسية. على وجه التحديد، يتم تنزيل النماذج المدربة مسبقًا من tf.keras.applications.
بالإضافة إلى التصنيفات الجديدة، يوفر ImageNet-PI أيضًا بيانات وصفية حول عملية التعليق التوضيحي في شكل ثقة النماذج في تصنيفاتها ومعلومات إضافية حول كل نموذج.
للمزيد من المعلومات، انظر: ImageNet-PI
الصفحة الرئيسية : https://github.com/google-research-datasets/imagenet_pi/
شفرة المصدر :
tfds.datasets.imagenet_pi.Builderالإصدارات :
-
1.0.0(الافتراضي): الإصدار الأولي.
-
حجم التنزيل :
Unknown sizeحجم مجموعة البيانات :
Unknown sizeتعليمات التنزيل اليدوي : تتطلب مجموعة البيانات هذه تنزيل بيانات المصدر يدويًا إلى
download_config.manual_dir(المسار الافتراضي هو~/tensorflow_datasets/downloads/manual/):
يجب أن يحتوي المجلد manual_dir على ملفين: ILSVRC2012_img_train.tar و ILSVRC2012_img_val.tar. للحصول على رابط تحميل مجموعة البيانات، عليك التسجيل على الموقع http://www.image-net.org/download-images .التخزين المؤقت التلقائي ( الوثائق ): غير معروف
تقسيمات :
| ينقسم | أمثلة |
|---|
- بنية الميزات :
FeaturesDict({
'annotator_confidences': Tensor(shape=(16,), dtype=float32),
'annotator_labels': Tensor(shape=(16,), dtype=int64),
'clean_label': ClassLabel(shape=(), dtype=int64, num_classes=1000),
'file_name': Text(shape=(), dtype=string),
'image': Image(shape=(None, None, 3), dtype=uint8),
})
- توثيق الميزات :
| ميزة | فصل | شكل | نوع البيانات | وصف |
|---|---|---|---|---|
| قاموس الميزات | ||||
| ثقة المعلق | الموتر | (16,) | float32 | |
| تسميات المعلقين | الموتر | (16,) | int64 | |
| علامة نظيفة | ClassLabel | int64 | ||
| اسم الملف | نص | خيط | ||
| صورة | صورة | (لا شيء، لا شيء، 3) | uint8 |
المفاتيح الخاضعة للإشراف (انظر وثيقة
as_supervised):('image', 'annotator_labels')الشكل ( tfds.show_examples ): غير مدعوم.
أمثلة ( tfds.as_dataframe ): مفقود.
الاقتباس :
@inproceedings{tram,
author = {Mark Collier and
Rodolphe Jenatton and
Effrosyni Kokiopoulou and
Jesse Berent},
editor = {Kamalika Chaudhuri and
Stefanie Jegelka and
Le Song and
Csaba Szepesv{\'{a} }ri and
Gang Niu and
Sivan Sabato},
title = {Transfer and Marginalize: Explaining Away Label Noise with Privileged
Information},
booktitle = {International Conference on Machine Learning, {ICML} 2022, 17-23 July
2022, Baltimore, Maryland, {USA} },
series = {Proceedings of Machine Learning Research},
volume = {162},
pages = {4219--4237},
publisher = { {PMLR} },
year = {2022},
url = {https://proceedings.mlr.press/v162/collier22a.html},
}
@article{ILSVRC15,
Author = {Olga Russakovsky and Jia Deng and Hao Su and Jonathan Krause and Sanjeev Satheesh and Sean Ma and Zhiheng Huang and Andrej Karpathy and Aditya Khosla and Michael Bernstein and Alexander C. Berg and Li Fei-Fei},
Title = { {ImageNet Large Scale Visual Recognition Challenge} },
Year = {2015},
journal = {International Journal of Computer Vision (IJCV)},
doi = {10.1007/s11263-015-0816-y},
volume={115},
number={3},
pages={211-252}
}