ออมนิกล็อต

  • คำอธิบาย :

ชุดข้อมูล Omniglot สำหรับการเรียนรู้แบบช็อตเดียว ชุดข้อมูลนี้ประกอบด้วยอักขระที่เขียนด้วยลายมือที่แตกต่างกัน 1,623 ตัวจากตัวอักษรที่แตกต่างกัน 50 ตัว

แยก ตัวอย่าง
'small1' 2,720
'small2' 3,120
'test' 13,180
'train' 19,280
  • โครงสร้างคุณสมบัติ :
FeaturesDict({
   
'alphabet': ClassLabel(shape=(), dtype=int64, num_classes=50),
   
'alphabet_char_id': int64,
   
'image': Image(shape=(105, 105, 3), dtype=uint8),
   
'label': ClassLabel(shape=(), dtype=int64, num_classes=1623),
})
  • เอกสารคุณสมบัติ :
คุณสมบัติ ระดับ รูปร่าง ประเภทD คำอธิบาย
คุณสมบัติDict
ตัวอักษร ClassLabel int64
ตัวอักษร_char_id เทนเซอร์ int64
ภาพ ภาพ (105, 105, 3) uint8
ฉลาก ClassLabel int64

การแสดงภาพ

  • การอ้างอิง :
@article{lake2015human,
  title
={Human-level concept learning through probabilistic program induction},
  author
={Lake, Brenden M and Salakhutdinov, Ruslan and Tenenbaum, Joshua B},
  journal
={Science},
  volume
={350},
  number
={6266},
  pages
={1332--1338},
  year
={2015},
  publisher
={American Association for the Advancement of Science}
}