Skip to content

Dataset · m-a-p

FineFineWeb: domain-specific web corpus

m-a-p/FineFineWeb

Large English web corpus broken down by dozens of domains and iteration statistics for training and classification.

What's inside

The dataset contains English web texts and tabular metadata broken down by domain—from aerospace to agronomy. The card provides token and example volumes across three collection iterations and final values. According to metadata, the size ranges from 1 to 10 billion; tasks include text-classification and text-generation.

What you can do with it

  • Train domain text classification
  • Fine-tune LLMs on a domain web corpus
  • Gather a domain dataset for generation
Tasks
Text classification, Text generation
Languages
English
Size
1B–10B rows
Downloads this month
435.2K
Likes
244

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("m-a-p/FineFineWeb")