i_naturalist2021

  • विवरण :

iNaturalist डेटासेट 2021 में कुल 10,000 प्रजातियां शामिल हैं। पूर्ण प्रशिक्षण डेटासेट में लगभग 2.7M चित्र हैं। डेटासेट को अधिक सुलभ बनाने के लिए हमने कुल 500K छवियों के लिए प्रति प्रजाति 50 उदाहरणों के साथ एक "मिनी" प्रशिक्षण डेटासेट भी बनाया है। पूरी ट्रेनिंग train स्प्लिट mini स्प्लिट के साथ ओवरलैप होती है। वैल सेट में प्रत्येक प्रजाति के लिए 10 सत्यापन छवियां (कुल 100K) शामिल हैं। public_test स्प्लिट (जमीनी सच्चाई लेबल के बिना) में कुल 500,000 परीक्षण छवियां हैं।

  • होमपेज : https://github.com/visipedia/inat_comp/tree/master/2021

  • स्रोत कोड : tfds.image_classification.i_naturalist2021.INaturalist2021

  • संस्करण :

    • 1.0.0 : प्रारंभिक रिलीज।
    • 2.0.0 : अपडेट: क्लास इंडेक्स JSON ट्रेन फाइल में ऑर्डर का पालन करते हैं।
    • 2.0.1 (डिफ़ॉल्ट) : अपडेट: JSON फ़ाइलों में प्रदान की गई उदाहरण आईडी शामिल करें।
  • डाउनलोड आकार : 316.54 GiB

  • डेटासेट का आकार : 318.45 GiB

  • ऑटो-कैश्ड ( दस्तावेज़ीकरण ): नहीं

  • विभाजन :

विभाजित करना उदाहरण
'mini' 500,000
'test' 500,000
'train' 2,686,843
'val' 100,000
  • फ़ीचर संरचना :
FeaturesDict({
    'file_id': Text(shape=(), dtype=string),
    'id': Scalar(shape=(), dtype=int64),
    'image': Image(shape=(None, None, 3), dtype=uint8),
    'label': ClassLabel(shape=(), dtype=int64, num_classes=10000),
    'supercategory': ClassLabel(shape=(), dtype=int64, num_classes=11),
})
  • फ़ीचर दस्तावेज़ीकरण :
विशेषता कक्षा आकार डीटाइप विवरण
विशेषताएं डिक्ट
file_id मूलपाठ डोरी
पहचान अदिश int64
छवि छवि (कोई नहीं, कोई नहीं, 3) uint8
लेबल क्लासलेबल int64
superश्रेणी क्लासलेबल int64

VISUALIZATION

  • उद्धरण :
\
@misc{inaturalist21,
    Howpublished = {~\url{https://github.com/visipedia/inat_comp/tree/master/2021} },
    Title = { {iNaturalist} 2021 competition dataset.},
    Year = {2021},
    key = { {iNaturalist} 2021 competition dataset},
    }