Skip to content

Releases: google-deepmind/rlax

RLax 0.1.6

29 Jun 15:03
Compare
Choose a tag to compare

What's Changed

  • Bump ipython from 7.16.1 to 8.10.0 in /requirements by @dependabot in #116
  • Fix KL constraint loss to ensure lagrange multiplier is always positive. by @copybara-service in #123
  • Drop python 3.7 and 3.8

New Contributors

Full Changelog: v0.1.5...v0.1.6

RLax 0.1.5

09 Jan 13:23
Compare
Choose a tag to compare

What's Changed

  • Replace for-loop in extract_subsequences with single indexing operation. by @copybara-service in #98
  • Replace O(n^2) iterative insert with linear append + reverse. by @copybara-service in #99
  • Expose utilities for constructing and learning from policy targets. by @copybara-service in #106
  • Add support for disabling stop_gradients on targets (as in other rlax losses). by @copybara-service in #107
  • [rlax] Update jax and numpy requirements for RLax. by @copybara-service in #114
  • Release new RLax version. by @copybara-service in #115

Full Changelog: v0.1.4...v0.1.5

RLax 0.1.4

15 Aug 07:29
Compare
Choose a tag to compare

What's Changed

  • rlax: Replace rlax categorical cross entropy computation with distrax components. by @copybara-service in #57
  • Bugfix to quantile_expected_sarsa. by @copybara-service in #63
  • Update Jinja2 versioning to avoid Sphinx failures. by @copybara-service in #66
  • Add test for squashed gaussian in rlax distributions. by @copybara-service in #68
  • Update squashed gaussian distribution in rlax for prob and logprob to numerically match distrax's implementation. by @copybara-service in #69
  • Migrate RLax squashed gaussian to use Distrax. Explicitly broadcast shapes in Distrax scalar affine to avoid rank promotion errors. by @copybara-service in #70
  • Add a particular pair of transforms used by muzero that combine a non linear squashing function with a reparametrisation of the scalar as linear combination of two hot values in a discrete suppport. by @copybara-service in #73
  • Support Array lambda_ in Vtrace. by @copybara-service in #71
  • Send deprecation warning for rlax.distributions in favor of using distrax. by @copybara-service in #74
  • Send deprecation warning for rlax nested_updates in favor of using optax. by @copybara-service in #75
  • Move usages of soon to be deprecated rlax.periodic_update to optax.periodic_update. by @copybara-service in #77
  • Add a pair of transforms where the scalar values are reparametrised as the linear combination of two-hot values on a non-linearly spaced discrete support. by @copybara-service in #78
  • Add moving averages helpers to rlax. by @copybara-service in #79
  • Update .pylintrc by @copybara-service in #80
  • Add utilities to extract overlapping subsequences from trajectories. by @copybara-service in #81
  • Minor edits to moving averages. by @copybara-service in #83
  • Add utilities to support interruptions. by @copybara-service in #84
  • Create new version 0.1.3 of RLax. by @copybara-service in #86
  • Remove incremental_update from rlax: all usages ported to optax.incremental_update by @copybara-service in #85
  • Pin numpy version <1.23 until new jax version is released, fixing bug that makes mpo_ops_test fail. by @copybara-service in #92
  • Fix a bug in tree_split_leaves(): squeeze the right axis in case of keepdim=False. by @copybara-service in #94
  • Fix max_start_idx argument. by @copybara-service in #87
  • Release a new rlax verison. by @copybara-service in #96

Full Changelog: v0.1.2...v0.1.4

RLax 0.1.2

24 Feb 15:27
Compare
Choose a tag to compare

What's Changed

  • Fix arg docstring for rho_tm1 and internal computations based on it to reflect time tm1 instead of t. by @copybara-service in #43
  • Add Sphinx build to CI test, point to documentation in README, and fix issues in doc strings that were causing CI test to fail. by @copybara-service in #46
  • Remove usages of apply_rng=True from Haiku code. by @copybara-service in #47
  • Add KNN Query to RLax public API. by @copybara-service in #44
  • Change RLax citation to Jax Ecosystem citation. by @copybara-service in #48
  • Update requirements and allow new versions of JAX. by @copybara-service in #50
  • Remove the old venv directory before testing the package. by @copybara-service in #52
  • Move decoupled_multivariate_normal_kl_divergence out of distributions.py by @copybara-service in #55
  • Use distrax distributions in epsilon_softmax. by @copybara-service in #59

Full Changelog: v0.1.1...v0.1.2

RLax 0.1.1

19 Nov 12:34
Compare
Choose a tag to compare

What's Changed

  • Drop python 3.6 support and release a new version. by @copybara-service in #42

Full Changelog: v0.1.0...v0.1.1

RLax 0.0.5

18 Nov 19:47
Compare
Choose a tag to compare

What's Changed

  • Fix failing copybara lint errors. by @copybara-service in #24
  • Add tests for clipped_entropy_softmax distribution and fix improperly negated clipped entropy. by @copybara-service in #21
  • Add tests for multivariate_normal_kl_divergence & kl functions in gaussian_diagonal. by @copybara-service in #22
  • Migrate RLax distributions to use distrax. by @copybara-service in #20
  • Re-allow rlax gaussian diagonal to work with scalar sigma. by @copybara-service in #25
  • Fixes bug in kl calculation of gaussian_diagonal by @copybara-service in #26
  • [JAX] Replace uses of deprecated jax.ops.index_update(x, idx, y) APIs with their up-to-date, more succinct equivalent x.at[idx].set(y). by @copybara-service in #33
  • [JAX] Increase numerical tolerances of tests in preparation for an XLA:CPU vectorization change. by @copybara-service in #34
  • Fix performance issue in simple DQN example. by @copybara-service in #32
  • Add test.sh for launching CI tests on a local machine. by @copybara-service in #35
  • Iterate over Python range instead jnp.arange. by @copybara-service in #29
  • fix kl argument order for gaussians by @akssri-sai in #15
  • Freeze the latest compatible JAX version. by @copybara-service in #36
  • Internal change. by @copybara-service in #38

Full Changelog: v0.0.4...v0.0.5

RLax 0.0.4

08 Jul 18:31
Compare
Choose a tag to compare

Note: this is a first GitHub release of RLax. It includes all changes since the repo was created.

Changelog

Unreleased

Full Changelog

Fixed bugs:

  • can not find setup.py for pip install #2

Closed issues:

  • Does discount = 0 mean "terminal" state by design? #17
  • vtrace uses lax.scan? #16
  • rlax is broken on Python 3.9 #13
  • missing library: import optax #8
  • Documentation and Examples #5

Merged pull requests:

* This Changelog was automatically generated by github_changelog_generator