README.md

metadata

license: apache-2.0
dataset_info:
  features:
    - name: title
      dtype: string
    - name: author
      dtype: string
    - name: dynasty
      dtype: string
    - name: theme
      dtype: string
    - name: section
      dtype: string
    - name: content
      list:
        - name: paragraphs
          sequence: string
        - name: subchapter
          dtype: string
    - name: appreciation
      dtype: string
    - name: rhythmic
      dtype: string
    - name: tags
      sequence: string
  splits:
    - name: train
      num_bytes: 100641630
      num_examples: 346320
  download_size: 66721235
  dataset_size: 100641630
language:
  - zh
tags:
  - poem
size_categories:
  - 100K<n<1M

数据格式定义

from pydantic import BaseModel, Field

class PoemValidator(BaseModel):
    title: str = Field(..., description="诗名")
    author: str = Field(..., description="作者")
    dynasty: str = Field(..., description="朝代")
    theme: str = Field(..., description="主题")
    section: str = Field(..., description="节名")
    content: Union[list[dict[str, Union[str, list[str]]]]] = Field(..., description="诗内容")
    appreciation: str = Field(..., description="赏析")
    rhythmic: str = Field(..., description="韵律")
    tags: list[str] = Field(..., description="标签")

加载数据集

from datasets import load_dataset

ds = load_dataset("ticoAg/cotinus-poem")