Skip to content

Dataset · JetBrains

JetBrains dataset for next code edit prediction

JetBrains/nessie-set

Real GitHub commit examples for training models to predict the next code change.

What's inside

Contains between 100k and 1M examples in text and tabular formats collected from commit histories of open-source GitHub repositories. Each entry includes the current file state and recent change history, from which the next developer edit must be predicted (including deletions and changes anywhere in the file).

What you can do with it

  • Training AI assistants for Next Edit Suggestion
  • Fine-tuning code LLMs on real file editing trajectories
Tasks
Text generation
Languages
Code
Size
100K–1M rows
Downloads this month
543
Likes
8

How to load it

from datasets import load_dataset

ds = load_dataset("JetBrains/nessie-set")