JuliaRL_A2C_CartPole experiment hook TotalRewardPerEpisode not work about reinforcementlearningzoo.jl HOT 2 CLOSED

This works for me

id = "JuliaRL_A2C_CartPole"
e = Experiment(id)
agent = e.policy
env = e.env
stop_condition = e.stop_condition
hook = TotalBatchRewardPerEpisode(length(e.env)) # Keep track of reward for each env
run(agent, env, stop_condition, hook)
rewards = hook.rewards

from reinforcementlearningzoo.jl.

The TotalRewardPerEpisode assumes a scalar reward from the reward(env) call, see

https://github.com/JuliaReinforcementLearning/ReinforcementLearningCore.jl/blob/c14ec876311408096a677f1b998e307cafff4dc4/src/core/hooks.jl#L124

Just running run(e) after defining the experiment works and will collect the reward for each environment, but since this is running multiple environments, see

https://github.com/JuliaReinforcementLearning/ReinforcementLearningZoo.jl/blob/master/src/experiments/rl_envs/JuliaRL_A2C_CartPole.jl

we instead use a TotalBatchRewardPerEpisode that takes an array of rewards from each environment, see

https://github.com/JuliaReinforcementLearning/ReinforcementLearningCore.jl/blob/c14ec876311408096a677f1b998e307cafff4dc4/src/core/hooks.jl#L152

so that is why it did not work.

from reinforcementlearningzoo.jl.