- বর্ণনা :
এক-শট শেখার জন্য অমনিগ্লট ডেটা সেট। এই ডেটাসেটে 50টি ভিন্ন বর্ণমালা থেকে 1623টি ভিন্ন হস্তলিখিত অক্ষর রয়েছে।
অতিরিক্ত ডকুমেন্টেশন : কোড সহ কাগজপত্রে অন্বেষণ করুন
সোর্স কোড :
tfds.image_classification.Omniglotসংস্করণ :
-
3.0.0(ডিফল্ট): নতুন স্প্লিট API ( https://tensorflow.org/datasets/splits )
-
ডাউনলোড সাইজ :
17.95 MiBডেটাসেটের আকার :
12.29 MiBস্বয়ংক্রিয় ক্যাশে ( ডকুমেন্টেশন ): হ্যাঁ
বিভাজন :
| বিভক্ত | উদাহরণ |
|---|---|
'small1' | 2,720 |
'small2' | 3,120 |
'test' | 13,180 |
'train' | 19,280 |
- বৈশিষ্ট্য গঠন :
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),
})
- বৈশিষ্ট্য ডকুমেন্টেশন :
| বৈশিষ্ট্য | ক্লাস | আকৃতি | ডিটাইপ | বর্ণনা |
|---|---|---|---|---|
| ফিচারসডিক্ট | ||||
| বর্ণমালা | ক্লাসলেবেল | int64 | ||
| alphabet_char_id | টেনসর | int64 | ||
| ইমেজ | ছবি | (105, 105, 3) | uint8 | |
| লেবেল | ক্লাসলেবেল | int64 |
তত্ত্বাবধান করা কী (দেখুন
as_superviseddoc ):('image', 'label')চিত্র ( tfds.show_examples ):

- উদাহরণ ( tfds.as_dataframe ):
- উদ্ধৃতি :
@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}
}