Skip to content

skrl-v2.1.1

Latest
Latest

Choose a tag to compare

View all tags
@Toni-SM Toni-SM released this 21 Oct 08:42
· 1 commit to main since this release
Immutable release. Only release title and notes can be modified.
2.1.1
21eba2b
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

[2.1.1] - 2026-10-20

Changed (breaking changes)

  • Update the Isaac Lab environment loader to Isaac Lab 3. Older versions of Isaac Lab are not supported

Fixed

  • Fix Warp PPO agent's update not preprocessing sampled observations/states after the first learning epoch
  • Fix Warp Gaussian model sampling and log standard deviation gradients
  • Fix sequential trainer's single-environment multi-agent reset when agents finish/leave the episode
  • Fix JAX and Warp Ornstein-Uhlenbeck noise's mean reversion and innovations sampled with twice the standard deviation
  • Fix JAX and Warp Gaussian noise sampling with twice the standard deviation
  • Fix PyTorch multivariate Gaussian model using the variance instead of the standard deviation as covariance factor
  • Fix PyTorch TRPO agent's KL divergence using the wrong sign for the log standard deviation term
  • Fix Gym space conversion dropping the start of Discrete spaces and the dtype of MultiDiscrete spaces

Removed (breaking changes)

  • Remove the random_timesteps and learning_starts configuration fields from on-policy agents/multi-agents
Assets 3
Loading