wiki_split

참조:

• 암호
• 허깅페이스

다음 명령을 사용하여 TFDS에서 이 데이터세트를 로드합니다.

ds = tfds.load('huggingface:wiki_split')

• 설명 :

One million English sentences, each split into two sentences that together preserve the original meaning, extracted from Wikipedia 
Google's WikiSplit dataset was constructed automatically from the publicly available Wikipedia revision history. Although 
the dataset contains some inherent noise, it can serve as valuable training data for models that split or merge sentences.

• 라이선스 : 알려진 라이선스 없음
• 버전 : 0.1.0
• 분할 :

• 특징 :

{
    "complex_sentence": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_sentence_1": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "simple_sentence_2": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    }
}