• I have marked all applicable categories:
  • exception-raising bug
  • RL algorithm bug
  • documentation request (i.e. "X is missing from the documentation.")
  • new feature request
• I have visited the source website
• I have searched through the issue tracker for duplicates
• I have mentioned version numbers, operating system and environment, where applicable:

  import tianshou, torch, sys
  print(tianshou.__version__, torch.__version__, sys.version, sys.platform)

In the DQN implementation, a policy network is required as an input to the DQN policy constructor. This is strange because DQN should only learn the Q-networks, and the policy should be completely determined by the learned Q-values (and the epsilon). I do not understand why a policy network is required as an input, but not a Q-network? Is it a particular variant of DQN? If so, I hope the documentation can be clear on it. And what if I want to use my own architecture for the Q-networks?