A problem about one game in ALE cannot be trained

Kaixhin / Rainbow

Rainbow: Combining Improvements in Deep Reinforcement Learning

MIT License

1.56k stars 282 forks source link

A problem about one game in ALE cannot be trained #83

Open Hugh-Cai opened 3 years ago

Hugh-Cai commented 3 years ago

Hi, Kai! I find an issue which happens when I set the game "defender" as the environment. It only displays hyper-parameter setting "args", and however, any training results aren't output, not as the same as other games.

Thanks!

Kaixhin commented 3 years ago

Can confirm. Something goes wrong during the environment creation, but that's atari_py failing on loadROM, so I suggest raising an issue in that repo. The following lines are enough to cause Python to hang for me:

import atari_py
ale = atari_py.ALEInterface()
ale.loadROM(atari_py.get_game_path('defender'))