Skip to content

Dataset · common-pile

Large Stack Exchange Q&A corpus

common-pile/stackexchange

Tens of millions of English discussions from various Stack Exchange communities for text generation training.

What's inside

The dataset contains 10 to 100 million English text entries from thematic Q&A communities across the Stack Exchange network (including closed sections like windowsphone). Materials are gathered from site dumps of the platform, whose publication in the Internet Archive was discontinued in July 2024.

What you can do with it

  • Pretraining and fine-tuning language models on technical and expert topics
  • Prepare instruction data (SFT) in a question-and-detailed-answer format
  • Populate knowledge bases for RAG systems on programming and related topics
Tasks
Text generation
Languages
English
Size
10M–100M rows
Downloads this month
9.2K
Likes
11

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("common-pile/stackexchange")