[Question] How to specify environment's parameters from command line arguments when playing/testing a trained policy with rl_games?

[Antonello-Scaldaferri-adi]

When running a training for the Peg Insertion task with rl_games, I can specify the environment's parameters from command line arguments:

./isaaclab.sh -p scripts/reinforcement_learning/rl_games/train.py --task Isaac-Factory-PegInsert-Direct-v0 --num_envs 4 --headless env.task.fixed_asset_init_pos_noise='[0.15, 0.15, 0.05]'

When running the play script to test the trained policy:

./isaaclab.sh -p scripts/reinforcement_learning/rl_games/play.py --task Isaac-Factory-PegInsert-Direct-v0 --num_envs 4 env.task.fixed_asset_init_pos_noise='[0.15, 0.15, 0.05]'

I get:

play.py: error: unrecognized arguments: env.task.fixed_asset_init_pos_noise=[0.15, 0.15, 0.05]

Am I doing it incorrectly, or is it possible to change the environment's parameters from command line arguments for training only and not for playing/testing?

Thanks a lot in advance for any clarification.