시각화 : Know Your Data에서 탐색
설명 :
dSprites는 6개의 실측 독립 잠재 요인에서 절차적으로 생성된 2D 모양의 데이터 세트입니다. 이러한 요소는 스프라이트의 color , shape , scale , rotation , x 및 y 위치입니다.
이러한 잠재성의 모든 가능한 조합은 정확히 한 번만 존재하며 N = 737280개의 총 이미지를 생성합니다.
잠재 요인 값
- 색상: 화이트
- 모양: 사각형, 타원형, 심장
- 배율: [0.5, 1]에서 선형 간격으로 6개 값
- 방향: [0, 2파이]의 40개 값
- 위치 X: [0, 1]의 32개 값
- 위치 Y: [0, 1]의 32개 값
우리는 한 번에 하나의 잠재성을 변경하고(위치 Y에서 시작하여 위치 X 등) 고정된 순서로 이미지를 순차적으로 저장했습니다. 따라서 첫 번째 차원의 순서는 고정되어 해당 이미지에 해당하는 잠재 값으로 다시 매핑할 수 있습니다.
우리는 모든 픽셀 출력이 서로 다른지 확인하면서 가장 작은 단계 변화를 갖도록 의도적으로 잠재 값을 선택했습니다. 노이즈가 추가되지 않았습니다.
소스 코드 :
tfds.datasets.dsprites.Builder
버전 :
-
2.0.0
(기본값): 새로운 분할 API( https://tensorflow.org/datasets/splits ) -
2.1.0
: 릴리스 노트가 없습니다.
-
다운로드 크기 :
26.73 MiB
데이터세트 크기 :
Unknown size
자동 캐시 됨( 문서 ): 알 수 없음
분할 :
나뉘다 | 예 |
---|---|
'train' | 737,280 |
- 기능 구조 :
FeaturesDict({
'image': Image(shape=(64, 64, 1), dtype=uint8),
'label_orientation': ClassLabel(shape=(), dtype=int64, num_classes=40),
'label_scale': ClassLabel(shape=(), dtype=int64, num_classes=6),
'label_shape': ClassLabel(shape=(), dtype=int64, num_classes=3),
'label_x_position': ClassLabel(shape=(), dtype=int64, num_classes=32),
'label_y_position': ClassLabel(shape=(), dtype=int64, num_classes=32),
'value_orientation': float32,
'value_scale': float32,
'value_shape': float32,
'value_x_position': float32,
'value_y_position': float32,
})
- 기능 문서 :
특징 | 수업 | 모양 | D타입 | 설명 |
---|---|---|---|---|
풍모Dict | ||||
영상 | 영상 | (64, 64, 1) | uint8 | |
label_orientation | 클래스 레이블 | int64 | ||
label_scale | 클래스 레이블 | int64 | ||
label_shape | 클래스 레이블 | int64 | ||
label_x_위치 | 클래스 레이블 | int64 | ||
label_y_position | 클래스 레이블 | int64 | ||
value_orientation | 텐서 | float32 | ||
가치 척도 | 텐서 | float32 | ||
value_shape | 텐서 | float32 | ||
값_x_위치 | 텐서 | float32 | ||
값_y_위치 | 텐서 | float32 |
감독된 키 (
as_supervised
문서 참조):None
그림 ( tfds.show_examples ):
- 예 ( tfds.as_dataframe ):
- 인용 :
@misc{dsprites17,
author = {Loic Matthey and Irina Higgins and Demis Hassabis and Alexander Lerchner},
title = {dSprites: Disentanglement testing Sprites dataset},
howpublished= {https://github.com/deepmind/dsprites-dataset/},
year = "2017",
}