v0.5.0

Link to Release PR - #238

🛠️ Core Framework Improvements

Custom Repertoire Integration (#222)

• Allow MAPElites to be built with any repertoire by passing custom repertoire init functions
• Default behavior maintained with MapElitesRepertoire.init
• Enhanced flexibility for custom repertoire implementations

MAPElites Repertoire Enhancement (#211)

• Enhanced MAPElitesRepertoire to store any kind of information through improved extra_scores handling
• Standardized extra scores management across all repertoire classes with keys_extra_scores parameter
• Added filter_extra_scores method to filter extra scores based on keys
• Moved observations from dedicated fields to extra_scores storage

Ask/Tell Interface (#221)

• New user-facing ask/tell interface for more flexible algorithm interaction
• Added init_ask_tell function for initialization without scoring functions
• Refactored update functions to use ask/tell internally, reducing code duplication

🔄 Major Refactoring & Architecture Changes

Selector System Overhaul (#209)

• New unified selection interface through the Selector abstraction for improved modularity and extensibility
• Introduced UniformSelector and MOMEUniformSelector implementations
• Selectors now return full sub-repertoires instead of just genotypes, eliminating the need for multiple sampling functions
• Default selection behavior integrated into repertoire classes

Brax Environment Structure Reorganization (#236)

• Complete restructuring of Brax-related code from /environments to /tasks/brax with proper versioning
• Added Brax v2 support alongside existing Brax v1
• Improved environment handling and version management

Code Organization (#200, #204)

• Moved cmaes.py from core to baselines for better module organization

🔧 Code Quality & Standards

Terminology Standardization (#190)

• Unified descriptor naming from mixed "bd", "BD", "behavior_descriptor" to consistent "descriptor"
• Stopped returning random keys following JAX best practices
• Renamed all random_key parameters to key
• Removed deprecated Brax scoring functions
• Standardized unused parameter naming with underscore prefix

JIT Compilation Cleanup (#232)

• Removed individual jax.jit decorators from internal functions for more user coding flexibility
• Improved function signatures

JAX Modernization (#195)

• Updated to new-style JAX random keys (jax.random.key instead of jax.random.PRNGKey)
• Full compliance with JAX JEP 9263 typed keys specification

⚡ Performance & Memory Optimizations

VRAM Optimization (#224)

• Significant memory improvements in PGA and DCRL algorithms
• Off-spring PG mutations now share mini-batches, reducing memory usage
• Step-by-step transition sampling in both "emit_pg" and "state_update"
• Simplified function structure and combined training/update operations
• New scan_actor_critic_training function for efficient Actor-Critic loops

🔍 Quality Assurance

Automated Spell Checking (#194)

• Added codespell pre-commit hook to catch typos automatically
• Fixed existing typos throughout codebase
• Updated pre-commit configuration with latest package versions

Updated example notebooks with better parameter handling for NSGA-II and SPEA2 (#231)

• Fixed batch size parameters to use population_size by default for genetic algorithms

📊 Metrics & Analysis Features

Uncertainty Domain Support (#186)

• New reevaluation functions in qdax/utils/uncertain_metrics.py
• Compute corrected archives for uncertain domains
• Comprehensive test coverage for uncertainty metrics

Enhanced Visualization (#233)

• Custom x-axis labels in plot_map_elites_results
• Defaults to "Environment steps" with backward compatibility

Metrics Initialization (#214)

• Return metrics after repertoire initialization for better monitoring

🧹 Cleanup & Maintenance

Simplified Interface (#225)

• Removed save and load methods from repertoire classes
• Users can now use custom serialization methods like pickle or orbax
• Reduced maintenance burden for new repertoire types

Package Configuration Modernization (#213)

• Migrated from setup.py to pyproject.toml for modern Python packaging standards

Fixed indexing in UnstructuredRepertoire (#185)

• better handling of NaN descriptors

Legacy spring support (#226)

• Legacy spring support for Brax v1 with backward compatibility boolean flag

v0.4.1

Full Changelog: v0.4.0...v0.4.1

What's Changed

• Fix #139: ensure MOME works for genotypes that are not arrays by @hannah-jan in #199
• Hotfix: remove useless dependencies (including gym dependency) by @Lookatator in #210

Full Changelog: v0.4.0...v0.4.1

v0.4.0

Updates and New Functionalities

• Add DCRL-ME algorithm - #167
• Upgrade Library Versions and Python Version - #187

Bug Fixes

• Fix the Pareto Dominance definition to account for solutions which have the same fitness values along one axis. - #174

Breaking Changes

• qdax/types.py has been renamed to qdax/custom_types.py
• Emitter.emit has a new output extra_info that is similar to the extra_scores of the scoring function, and that enables to pass information from the emit step to the state_update (necessary for DCRL-ME).
• Emitter.init now also takes as input a repertoire, as well as fitnesses, descriptors and extra_scores.
• init_genotypes have been replaced with genotypes in the arguments of Emitter.init

Documentation

• Clarify contribution process - #171

Full Changelog: v0.3.0...v0.4.0

v0.3.0

What's Changed

• Uniform sac network sizes, by @limbryan in #145
• Remove singularity tools and update haiku version, by @limbryan in #147
• Add MAP-Elites Low-Spread, by @btjanaka in #152
• Add MAP-Elites Low-Spread (#152), by @Lookatator in #154
• Fix spelling errors, by @btjanaka in #150
• Use brax.v1 and update requirements, by @DBraun in #156
• Add sampling extractors, by @manon-but-yes in #158
• Update the paper to cite, by @felixchalumeau in #161
• Fix dependencies in requirements.txt and setup.py, by @Lookatator in #165
• Fix imports jumpy and formatting issues, by @Lookatator in #163
• Add AURORA and PGA-AURORA, by @felixchalumeau in #144
• Fixing colab link in CMA-ME notebook, by @felixchalumeau in #172

New Contributors

• @btjanaka made their first contribution in #152
• @DBraun made their first contribution in #156

Full Changelog: v0.2.3...v0.3.0

v0.2.3

What's changed

• Fix dependencies in notebook examples

v0.2.2

V0.2.2

What's Changed

• minimal support for jumanji #136
• PBT and PBT-ME #143
• Humanoid Trap #143

Full Changelog: v0.2.1...v0.2.2

v0.2.1

Small release to fix a dependency that was preventing the installation of qdax through pypi. The issue was related to scikit-learn installation. See #130

v0.2.0

Moving to v0.2.0 with several new algorithms, fixes of current implementations, a whole new set of tasks, a more general API and some dependencies update to stay in phase with the Jax community 🚀 ⏩ ⏩

• fix(envs): order of wrappers to ensure update of state descriptor when using fixed init state (#128)
• fix(doc): add colab links, missing doc, update version (#125)
• feat!(repertoire): optional extra-scores for repertoire addition (#118)
• fix(jit): avoid consecutive jits of same method in for loops (#122)
• fix(docker): fix run-image docker stage (#121)
• feat(algorithms): add MAP-Elites distributed on multiple devices (#117)
• fix(test): inverse fitness and desc names in sampling test (#119)
• feat(github): add GitHub template for PR (#120)
• feat(algorithms): add QDPG emitter + refactor PGAME (#110)
• feat(algorithms): add CMA-ME, fix CMA-ES and CMA-MEGA (#86)
• fix(mees): add batch size property (#114)
• feat(algorithms): Add Multi-Emitter (#90)
• fix: reset_based scoring in brax_env default task (#109)
• feat(algorithms): add ME-ES to QDax (#81)
• fix(examples): brax version in colab examples (#108)
• fix(docs): avoid using flax 0.6.2 in setup (#112)
• feat(envs): wrapper for fixed initial state of environments (#92)
• fix(style): mypy issue in controller training
• fix: optimizer state reinitialization for PG variations (#104)
• fix: add update policy delay in PG emitter
• fix(pointmaze): scale after the clip of actions (#101)
• hotfix(images): re-add deleted logos to the repo
• docs: add caveats and logo (#99)
• feat: Default Scoring Functions for Sphere, Rastrigin, Arm, Brax environments, Hypervolume functions and QD Suite (#73)
• chore: Update jax, brax and flax versions (fixes the jax.tree_util warnings) (#76)
• doc: Add remark for installing QDax with GPU support in README (#77)
• fix the replay buffer overflow issue (#75)
• fix: correct irrelevant factor 0.25 in td3 loss (#78)

We also welcome a new contributor: @maxencefaldor 👏

Well done to all the team for this new release 🎉 👩‍💻 👨‍💻 (@manon-but-yes @Lookatator @limbryan @ranzenTom @Aneoshun @Egiob @valentinmace @maxiallard @felixchalumeau)

v0.1.0

Finally moving to version 0.1.0 of QDax with several algorithms additions, new tools, enhanced documentation and several fixes 🚀

• Update notebooks to run on colab and update chex version (#61)
• Pin brax version + remove autoreload (#60)”
• Enhance the current documentation + gather baselines in separate folder (#57)
• Add example notebook for NSGA2 and SPEA2 (#58)
• Add API documentation, use README as home page (#56)
• Add NSGA2 and SPEA2 to QDax (#33)
• Add reset_based_scoring_function for stochastic environments (#31)
• Issue with batch size 1 (#52)
• Add colab badges to all example notebooks (#53)
• Add codecov to QDax for test coverage reports (#39)
• Update singularity pipeline and fix pre-commits (#45)
• Add plot_multidimensional_map_elites_grid function (#47)
• Extend functionality of compute_euclidean_centroids (#40)
• Repertoire addition for 1d bd (map elites and mome) (#41)
• Static argnames in _sample_in_masked_pareto_front (#49)
• Update ci trigger rules (#48)
• Use rngkey for centroids generation
• Run workflow on pull requests
• Use RNGKey for sampling CVT centroids and for KMeans
• Fix mome emitter state update
• Add MOME to QDax (#27)
• Fix steps incrementation in pointmaze (#36)
• Upgrade requirements for numpy and jax (#34)
• Add CMA MEGA (#25) - (API change!)
• Add SMERL (DIAYN+DADS) (#22)
• Add DADS (#21)
• Add OMG-MEGA (#24)
• Add DIAYN (#20) and refactor SAC
• Add SAC algorithm (#16)
• Add TD3 algorithm (#14)
• Fix issue in behavior descriptor evaluation (#13)