Dataset

How Fast Do You Forget? 13M Duolingo Sessions

Contains 12.85M real vocabulary practice sessions across 6 language courses, compiled into forgetting curves, word difficulty tables, and recall stats.

Kaggle ↗Jul 12, 2026

  • tabular
  • spaced-repetition
  • forgetting-curves
  • language-learning
cover
cover

What it is

12.85M Duolingo vocabulary sessions from 115k learners, structured into forgetting curves, word difficulty tables, and recall stats across 6 languages. ~12.85M vocabulary practice sessions from 115,222 Duolingo learners across 6 languages, structured as analysis-ready forgetting curves, per-word difficulty tables, and course-level recall statistics.

Key results

  • 12,854,226 vocabulary practice sessions
  • 115,222 Duolingo learners across 6 languages
  • Derived from a 379 MB raw trace over a 13-day 2013 window

Useful for modeling recall decay and estimating item difficulty in spaced-repetition systems.

One useful ML email a week.

Worth your time, not all of your time.

We track new ML work across papers, code, and data, then keep only the pieces worth a closer look. Selected, not collected.

Papers.Data.Code

The Weekly ML Signal

Selected, not collected.

One useful ML email a week: the papers, repos and datasets worth your attention.

Worth your time, not all of your time.