pemami4911/deep-rl

Issues

• image inputs

  #20 opened 6 years ago by xxchenchen
  1

• Actor network output increases to 1, TORCS, TF 1.0.0

  #11 opened 8 years ago by Amir-Ramezani
  1

• L2 weight decay for Q

  #9 opened 8 years ago by Anjum48
  5

• In tf.gradients, why -self.action_gradien is needed?

  #19 opened 6 years ago by GoingMyWay
  1

• why is DDPG so unstable?

  #16 opened 7 years ago by DontGiveUpEasily
  2

• converger issue

  #2 opened 8 years ago by fangthu
  10

• why ddpy.py line 99 self.action_gradient has -

  #10 opened 8 years ago by jintw
  1

• Using DDPG for Pong

  #6 opened 8 years ago by option-greek
  4

• Possible regression to do with batch normalization

  #18 opened 7 years ago by JanMatas
  1

• batch normalization not actually enabled?

  #17 opened 7 years ago by codekitchen
  1

• A problem about the DDPG

  #15 opened 7 years ago by geweihgg
  5

• Divide actor gradients by batch size?

  #13 opened 8 years ago by andrewLevy
  2

• There is an error... about Monitor

  #12 opened 8 years ago by wonchul-kim
  0

• Issues: self.action_grads = tf.gradients(self.out, self.action)

  #3 opened 8 years ago by cxfneo
  13

• UPGRADE: Compatibility with latest version of Tensorflow (0.12.0-rc1)

  #4 opened 8 years ago by pemami4911
  3

• Some questions!

  #5 opened 8 years ago by windweller
  1

• OSError: dlopen: cannot load any more object with static TLS

  #1 opened 8 years ago by gabrieledcjr
  4