Hi, I used PPO to train the agent and saved the policy using checkpoint. But when I restored the policy and using compute single action function to get action, the policy will output action that out of action spaces. So, I want to know if I did something wrong or there are problems in Rllib. By th…

Restored Policy gives action that is out of bound

arturn April 13, 2023, 11:06pm 2

Have a look at what I wrote over here.
Algorithm unsquashes actions.

Topic		Replies	Views
[rllib] Problem running compute_single_action from PPO restored checkpoint Checkpointing, Restoring	1	183	December 13, 2023
Prediction outside outside action space during inference	0	59	March 18, 2024
How to deploy a trained Ray RLlib PPO policy/model in multi-agent-case? RLlib	5	533	November 10, 2021
Inconsistent actions from Algorithm.compute_single_action RLlib	3	245	June 14, 2023
Compute_single_action(obs, state) of policy and algo: different performance Checkpointing, Restoring	1	470	April 13, 2023