Issues

• JAX and Warp OU noise use theta as retention instead of mean reversion

  Toni-SM/skrl#474 · Filter by author Afloat16 opened on Oct 8, 2026

• Memory rejects reuse of an existing multidimensional tensor

  Toni-SM/skrl#472 · Filter by author aalquwayfili opened on Oct 8, 2026

• RunningStandardScaler merges sample variance as population variance

  Toni-SM/skrl#470 · Filter by author Afloat16 opened on Oct 7, 2026

• PyTorch TRPO uses the wrong sign for the Gaussian KL log-scale term

  Toni-SM/skrl#468 · Filter by author Afloat16 opened on Oct 7, 2026

• Multivariate Gaussian uses variance as its Cholesky factor

  Toni-SM/skrl#466 · Filter by author Afloat16 opened on Oct 7, 2026

• SAC checkpoint loading omits the learned entropy temperature and cached coefficient

  Toni-SM/skrl#464 · Filter by author Afloat16 opened on Oct 7, 2026

• Gym conversion drops Discrete start and MultiDiscrete dtype

  Toni-SM/skrl#460 · Filter by author betacatsling opened on Sep 12, 2026

• sample_all omits transitions when memory size is not divisible by mini_batches

  Toni-SM/skrl#458 · Filter by author betacatsling opened on Sep 12, 2026

• Repeated sample_space calls return identical batches of random actions

  Toni-SM/skrl#453 · Filter by author betacatsling opened on Sep 11, 2026

• Policy/value learning-rate tuples are ignored by Torch A2C and PPO

  need-triage
  Toni-SM/skrl#451 · Filter by author KirillHit opened on Sep 3, 2026
  1

• a logic problem about should_reset

  need-triage
  Toni-SM/skrl#445 · Filter by author fishhide opened on May 21, 2026

• Support shared parameters between actor and critic modules

  need-triage
  Toni-SM/skrl#431 · Filter by author ngaylinn opened on Apr 25, 2026

• play.py in Isaac Lab

  bug
  Toni-SM/skrl#361 · Filter by author Yovan275 opened on Aug 24, 2025

• PPO with Random Steps

  bug
  Toni-SM/skrl#316 · Filter by author redzhepdx opened on May 2, 2025
  1

• DDPG: KeyError ' initial_scale'

  bug
  Toni-SM/skrl#310 · Filter by author pietrodardano opened on Apr 14, 2025

• Wrong shape of actions for vectorized Gymnasium environments with discrete action spaces

  bug
  Toni-SM/skrl#308 · Filter by author philsteg opened on Apr 7, 2025

• class directly calls the __init__ methods of its base classes (Model and GaussianMixin)

  bug
  Toni-SM/skrl#296 · Filter by author robot-feng opened on Mar 18, 2025

• PPO without Memory Buffer (Jax)

  bug
  Toni-SM/skrl#293 · Filter by author ShimBoi opened on Mar 17, 2025
  1

• Multi-Agent implementation only checks the first agent when computing the finished-episode info

  bug
  Toni-SM/skrl#289 · Filter by author Toni-SM opened on Mar 12, 2025
  2

  

• StepTrainer

  bug
  Toni-SM/skrl#288 · Filter by author raptoravis opened on Mar 12, 2025

• Loading weights from huggingface gives size mismatch.

  bug
  Toni-SM/skrl#223 · Filter by author utkarshp1161 opened on Nov 13, 2024

• Random action only samples from the first action space dimension

  bug
  Toni-SM/skrl#158 · Filter by author nikolaradulov opened on Jun 20, 2024
  

• Mean rewards are not calculated properly

  bug
  Toni-SM/skrl#157 · Filter by author nikolaradulov opened on Jun 20, 2024
  2