metadata

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
license: mit
pretty_name: shakespearefirstfolio

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:

It is a single file, which makes further processing difficult
It does not contain all of Shakespeare's works
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.

Usage

import datasets

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