바이오스

참조:

다음 명령을 사용하여 TFDS에서 이 데이터세트를 로드합니다.

ds = tfds.load('huggingface:biosses')
  • 설명 :
BIOSSES is a benchmark dataset for biomedical sentence similarity estimation. The dataset comprises 100 sentence pairs, in which each sentence was selected from the TAC (Text Analysis Conference) Biomedical Summarization Track Training Dataset containing articles from the biomedical domain. The sentence pairs were evaluated by five different human experts that judged their similarity and gave scores ranging from 0 (no relation) to 4 (equivalent).

  • 라이선스 : BIOSSES는 GNU Common Public License v.3.0의 조건에 따라 제공됩니다.

  • 버전 : 0.0.0

  • 분할 :

나뉘다 예
'train' 100
  • 특징 :
{
    "sentence1": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "sentence2": {
        "dtype": "string",
        "id": null,
        "_type": "Value"
    },
    "score": {
        "dtype": "float32",
        "id": null,
        "_type": "Value"
    }
}