Research map / Reinforcement learning / Visual and representation RL
QDax: A Library for Quality-Diversity and Population-based Algorithms with Hardware Acceleration
Felix Chalumeau, Bryan Lim, Raphaël Boige, Maxime Allard, Luca Grillotti, Manon Flageat, Valentin Macé, Guillaume Richard, Arthur Flajolet, Thomas Pierrot, Antoine Cully
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use QDax JAX implementation instead of the author PyTorch/multiprocessing implementation | PGA-MAP-Elites · runtime on affordable GPU | environment steps per hour: 2.8 × 10^7 -> 10^8 | Improved |
| use QDax JAX implementation instead of the authors' implementation | QD-PG · Ant Trap, 10^8 steps | wall-clock time: 15 hours -> 3 hours | Improved |
| use QDax on a single GPU instead of 1000 CPUs | ME-ES · Ant Maze | environment steps per hour: 6 × 10^8 -> 4 × 10^8 | No clear effect |
| add policy-gradient emitter (PGA-MAP-Elites vs MAP-Elites) | MAP-Elites · Brax tasks average | training steps in two hours: 1.66 × 10^9 -> 1.93 × 10^8 | Hurt |
| use population-based MAP-Elites instead of SMERL skill discovery | SMERL · Brax tasks average | training steps in two hours: 9.97 × 10^6 -> 1.66 × 10^9 | Improved |
| jit-compile with JAX on hardware accelerators | QDax algorithms · general runtime | speed-up: up to 10x | Improved |
Closest papers by meaning
- Proximal Policy Gradient Arborescence for Quality Diversity Reinforcement Learning (ICLR 2024 · similarity 0.82)
- XLand-MiniGrid: Scalable Meta-Reinforcement Learning Environments in JAX (NeurIPS 2024 · similarity 0.82)
- Quality-Diversity with Limited Resources (ICML 2024 · similarity 0.82)
- JaxMARL: Multi-Agent RL Environments and Algorithms in JAX (NeurIPS 2024 · similarity 0.82)
- Quality-Diversity Actor-Critic: Learning High-Performing and Diverse Behaviors via Value and Successor Features Critics (ICML 2024 · similarity 0.81)
- Behavioral Foundation Models for Quality Diversity (NeurIPS 2026 · similarity 0.81)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
