gvlassis/shakespearefirstfolio

---
dataset_info:
features:
- name: text
dtype: string
splits:
- name: train
num_bytes: 4077942
num_examples: 30
- name: validation
num_bytes: 245785
num_examples: 2
- name: test
num_bytes: 506679
num_examples: 4
download_size: 3073023
dataset_size: 4830406
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*
- split: test
path: data/test-*
task_categories:
- text-generation
language:
- en
tags:
- shakespeare
size_categories:
- n<1K
---
# shakespearefirstfolio

## About
🎭 Shakespeare's First Folio (a collection of 36 of Shakespeare's plays) as a Hugging Face dataset!

## Description
In 2015, Andrej Karpathy wrote a post called "The Unreasonable Effectiveness of Recurrent Neural Networks" in his blog. For the needs of this post, he created tinyshakespeare, a subset of Shakespeare's works in a single 40,000 lines file. Surprisingly, language models trained from scratch on this tiny dataset can produce samples that look very close to those written by Shakespeare himself.

Since then, tinyshakespeare has been the defacto dataset used as a first test while developing language models. Unfortunately, it has some problems:
1) It is a single file, which makes further processing difficult
2) It does not contain all of Shakespeare's works
3) It is not clear exactly what works and to what extend are included

This dataset tries to address these problems. It is ~4 times bigger than tinyshakespeare.

It was manually collected from [Folger Shakespeare Library](https://www.folger.edu/).

## Usage
import datasets

dataset = datasets.load_dataset("gvlassis/shakespearefirstfolio")

原始信息汇总

数据集概述

基本信息

• 名称: shakespearefirstfolio
• 任务类别: 文本生成
• 语言: 英语
• 标签: 莎士比亚
• 大小类别: n<1K

数据结构

• 特征:
  • text: 字符串类型

数据分割

• 训练集:
  • 示例数量: 30
  • 字节数: 4077942
• 验证集:
  • 示例数量: 2
  • 字节数: 245785
• 测试集:
  • 示例数量: 4
  • 字节数: 506679

数据大小

• 下载大小: 3073023字节
• 数据集大小: 4830406字节

配置文件

• 默认配置:
  • 训练数据路径: data/train-*
  • 验证数据路径: data/validation-*
  • 测试数据路径: data/test-*