mirror of
https://github.com/wassname/DeepRL.git
synced 2026-09-09 11:13:47 +08:00
Refactor DDPG
This commit is contained in:
@@ -46,15 +46,6 @@ class ContinuousAdvantageActorCritic:
|
||||
terminal = (terminal or (config.max_episode_length and steps >= config.max_episode_length))
|
||||
next_state = self.state_normalizer(next_state)
|
||||
|
||||
# if deterministic:
|
||||
# self.config.logger.scalar_summary('reward', reward, self.counter)
|
||||
# self.config.logger.histo_summary('std', std.data.numpy(), self.counter)
|
||||
# self.config.logger.histo_summary('mean', mean.data.numpy(), self.counter)
|
||||
# self.config.logger.histo_summary('action', action, self.counter)
|
||||
# self.config.logger.scalar_summary('steps', steps, self.counter)
|
||||
# self.config.logger.histo_summary('states', state, self.counter)
|
||||
# self.counter += 1
|
||||
|
||||
steps += 1
|
||||
total_reward += reward
|
||||
reward = np.asscalar(self.reward_normalizer(np.array([reward])))
|
||||
|
||||
Reference in New Issue
Block a user