中国诗歌飞花令

数据集文件元信息以及数据文件，请浏览“数据集文件”页面获取。

当前数据集卡片使用的是默认模版，数据集的贡献者未提供更加详细的数据集介绍，但是您可以通过如下GIT Clone命令，或者ModelScope SDK来下载数据集
# 唐诗飞花令数据集

唐诗飞花令数据集，适用于语言模型学习唐诗创作和接龙。

## 数据集特色

- **飞花令游戏**: 包含关键字飞花令、创作飞花令、接龙飞花令等多种形式
- **多任务学习**: 涵盖诗句引用、创作、接龙等不同任务
- **丰富关键字**: 包含春、花、秋、月等30个常见飞花令关键字

## 数据集配置

```yaml
license: Apache-2.0
task: text-generation
language: zh
tags:
- poetry
- tang-poetry
- feihualing
- chinese-game

configs:
- config_name: default
data_files:
- split: train
path: train.csv
- split: test
path: test.csv
```

## 字段说明

- `instruction`: 飞花令指令
- `input`: 输入内容/关键字
- `output`: 诗句输出
- `type`: 飞花令类型
- 飞花令-引用: 引用原诗中的诗句
- 飞花令-创作: 创作新诗句
- 飞花令-接龙: 诗句接龙
- 主题飞花令: 特定主题的诗句
- `keyword`: 飞花令关键字
- `source`: 诗句来源

## 使用示例

```python
from modelscope import MsDataset

# 加载数据集
dataset = MsDataset.load('your-username/唐诗飞花令数据集', split='train')

# 飞花令示例
for item in dataset:
if item['keyword'] == '月':
print(f"指令: {item['instruction']}")
print(f"输出: {item['output']}")
break
```

## 飞花令规则

飞花令是古代文人饮酒时的文字游戏，要求参与者轮流说出包含特定关键字的诗句。

## 数据来源

基于chinese-poetry数据库的唐诗数据整理。

#### 下载方法
:modelscope-code[]{type="sdk"}
:modelscope-code[]{type="git"}

Dataset file metadata and data files can be accessed via the "Dataset Files" page.

This dataset card uses the default template. The dataset contributors have not provided a more detailed introductory description, but you can download the dataset using the following Git Clone command or ModelScope SDK.

# Tang Poetry Feihualing Dataset

Tang Poetry Feihualing Dataset, designed for language models to learn Tang poetry composition and poetry continuation.

## Dataset Features
- **Feihualing Game**: Includes multiple forms such as keyword-based Feihualing, creative Feihualing, and poetry continuation Feihualing
- **Multi-task Learning**: Covers various tasks including poetry quotation, composition, and continuation
- **Rich Keywords**: Contains 30 common Feihualing keywords such as spring, flower, autumn, moon, etc.

## Dataset Configuration
yaml
license: Apache-2.0
task: text-generation
language: zh
tags:
- poetry
- tang-poetry
- feihualing
- chinese-game

configs:
- config_name: default
data_files:
- split: train
path: train.csv
- split: test
path: test.csv

## Field Explanations
- `instruction`: Feihualing instruction prompt
- `input`: Input content / target keyword
- `output`: Generated poetry verse
- `type`: Feihualing task type
- Feihualing-Quotation: Quote existing verses from original Tang poems
- Feihualing-Creation: Compose new original verses
- Feihualing-Continuation: Continue a given poetry verse
- Themed Feihualing: Verses for specific themed tasks
- `keyword`: Feihualing target keyword
- `source`: Source of the quoted or generated verse

## Usage Example
python
from modelscope import MsDataset

# Load the dataset
dataset = MsDataset.load('your-username/tang-poetry-feihualing-dataset', split='train')

# Feihualing demonstration
for item in dataset:
if item['keyword'] == '月':
print(f"Instruction: {item['instruction']}")
print(f"Generated Output: {item['output']}")
break

## Feihualing Rules
Feihualing is a traditional literary game among ancient Chinese scholars, typically played during wine banquets, where participants take turns reciting classical Chinese poetry verses containing a specified keyword.

## Data Source
Sorted and compiled from Tang poetry datasets sourced from the chinese-poetry database.

#### Download Methods
:modelscope-code[]{type="sdk"}
:modelscope-code[]{type="git"}

搜集汇总

背景与挑战

背景概述

该数据集为唐诗飞花令数据集，适用于语言模型学习唐诗创作与接续，包含多种飞花令形式和30个常见关键词。数据集基于chinese-poetry数据库的唐诗数据组织，涵盖训练和测试集，字段包括指令、输入、输出等类型。