Processor's def process_reward(self, reward), def process_step(self, observation, reward, done, info): wont return the value of reward for atari and retro.
toksis opened this issue · comments
toksis commented
Adding a print("Reward: ",reward) on this method process_reward(self, reward) or process_step(self, observation, reward, done, info): won't return a value. Always zero.
Error reproduction.
In the file dqn.atari.py in the example folder, paste this code. print("reward: ",reward).
Run the file.
Error: reward = 0.
This also happen in Retro.