RLlib

Debugging and performance tuning Steps to contribute to RLlib Configure Algorithm, Training, Evaluation, Scaling Repeated observation space Ray Tune stopping condition & comparisons Offline RL Checkpointing, Restoring

Topic	Replies	Views	Activity
RLlib beginners tutorial at this year's Ray Summit (June 22nd-24th)! RLlib	11	1741	June 29, 2021
How to pretrain a model with behavior cloning RLlib	14	5502	December 5, 2023
RLLIB not working with Tune with sample batch input RLlib	25	2689	October 4, 2022
_winapi.CreateProcess(executable, args, FileNotFoundError: [WinError 2] RLlib	9	18654	May 14, 2022
ValueError: Expected parameter logits (...) to satisfy the constraint IndependentConstraint(Real(), 1) RLlib	38	9174	October 14, 2024
RLlib Office Hours - Now open for signup RLlib	14	1369	July 1, 2025
[RLlib] Impossible actions RLlib	12	4139	May 11, 2022
Ray for Rapberry Pi, is possible? RLlib	30	4427	February 23, 2021
Observation_space not provided in PolicySpec RLlib	21	7498	February 7, 2023
Multi-Agent Training with Different Algorithms RLlib	24	3780	October 11, 2022
[RLlib] Visualise custom environment RLlib	18	4231	March 30, 2021
Reproducible training - setting seeds for all workers / environments RLlib	20	6247	May 24, 2023
Resume=True fails without useful error message RLlib	31	3388	September 26, 2022
Help debugging a memory leak in rllib RLlib	21	4065	September 25, 2022
Issues reproducing stable-baselines3 PPO performance with rllib RLlib	14	2718	March 16, 2022
Board game self-play PPO RLlib	15	4193	May 4, 2021
Issue creating custom action mask enviorment RLlib	14	2287	October 11, 2023
Deploying a learned policy under "explore=False / True" RLlib	9	1529	March 17, 2022
Compute_actions for Trajectory API RLlib	11	2461	February 10, 2022
RLlib, PyTorch and Mac M1 GPUs: No available node types can fulfill resource request RLlib	11	4322	February 29, 2024
Removing Algorithms from RLlib RLlib	10	1398	July 22, 2022
Meaning of episode_reward_mean RLlib	10	4386	September 21, 2023
Is any multi discrete action example for PPO or other algorithms? RLlib	9	4563	January 29, 2023
Use Policy_Trainer with TensorBoard RLlib	33	2467	November 13, 2021
RNN L2 weights regularization RLlib	41	2214	July 5, 2021
Observation space with multiple input RLlib	15	3530	December 10, 2021
Unable to restore fully trained checkpoint RLlib	19	3030	October 21, 2023
Issue with custom LSTMs RLlib	34	2246	February 26, 2023
Missing 'grad_gnorm' key in some `input_trees` after some training time RLlib	23	2364	January 29, 2023
How to define fcnet_hiddens size and number of layers in rllib tune? RLlib	18	2571	January 19, 2023
Apply preprocessor in custom model RLlib	19	2489	May 13, 2024
How do I set GPU affinity of workers RLlib	17	2573	April 23, 2021
RLLib Multiagent: Load only one policy from checkpoint & Compatibility of RLLib/Tune Checkpoints RLlib	9	3367	November 24, 2021
Compute/display actions from ray.tune RLlib	10	1725	March 30, 2021
Best way to have custom value state + LSTM RLlib	9	3151	April 10, 2022
Stacking callback objects [Solved. Code included.] RLlib	12	1538	April 30, 2021
Very slow gradient descent on remote workers RLlib	14	2518	June 8, 2021
Right way to use tuple action space RLlib	9	1662	September 24, 2021
Maximum recommended reward RLlib	18	2028	July 14, 2022
Accessing info dicts in postprocessing callback RLlib	10	1452	January 11, 2021
Custom RNN Model with Examples - why do they fail? RLlib	11	2395	May 5, 2022
Global optima with centralized critic (basic understanding) RLlib	10	2418	April 10, 2021
How to log Render to tensorboard? RLlib	9	2517	July 22, 2021
PPO trainer eating up memory RLlib	9	2431	April 2, 2021
Reward function not converging during training RLlib	14	1956	July 11, 2022
[Bug] Env must be one of the supported types: BaseEnv, gym.Env, MultiAgentEnv, VectorEnv, RemoteBaseEnv RLlib	10	2261	March 2, 2023
Error when running on GPU RLlib	9	2316	February 23, 2022
Get agent ID in multi-agent setting RLlib	16	1771	October 5, 2021
Memory Leak when training PPO on a single agent environment RLlib	15	1810	December 24, 2022
Playing the QMIX Two-step game on Ray RLlib	11	2090	October 18, 2022