Skip to content

Dataset · mbien

Largest recipe corpus

mbien/recipe_nlg

Over 2.23 million cleaned cooking recipes in English for text generation and search.

What's inside

Features 2,231,142 cooking recipes in English. According to the authors, the corpus has undergone more thorough processing and surpasses other open datasets in this field in volume.

What you can do with it

  • Training and fine-tuning recipe generation models
  • Building semantic search and text summarization systems
  • Fine-tuning language models on fill-mask tasks
Tasks
Text generation, Fill mask, Text retrieval, Summarization
Languages
English
Size
1M–10M rows
Downloads this month
513
Likes
54

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("mbien/recipe_nlg")