Repository Issue Activity (beta)

#562 GAE Bug for Envpool: Dummy Step Leak

#568 Misleading TensorBoard Command

#105 Numerical instability in C51

#563 Optional monitoring callback for reward-hacking / training-instability signals

#546 Bug in Usage of SumSegmentTree

#532 PPO Continues action

#518 DQN is no longer a single-file script

#553 Adding QR-DQN algorithm

#451 Docker image is out of date

#250 Adding Double DQN

#252 Adding unit tests

#429 [BUG] Different final epsilon and evaluation epsilon for Atari implementations

#545 KeyError: 'final_observation'. The return of envs.step(actions) changed in new gymnasium update (required for Mujoco v5 environments)

#525 PPO Hopper SOTA

#543 PQN's hyperparameters for classic control

#508 Logging of episodic returns in ppo implementations

#510 Proposal to Add PPO + Mamba SSM to CleanRL

#542 Potential IndentationError regressions in PR #539 (over-indented logging_step in JAX scripts)

#541 Issue #508: Episodic Return Logging Fix - Implementation Tracking

#540 CI: All JAX tests failing due to outdated JAX dependency stack

#533 Bug in cleanrl_utils/tuner.py

#534 [Bug/Feature] Support for vectorized infos structure in Gymnasium (missing final_info)

#442 Contributing PPO + Transformer-XL

#526 Support JAX environments

#460 Truncation not handled correctly when `optimize_memory_usage=True`