|
|
|
|
|
|
rewards_holder = tf.placeholder( |
|
|
|
shape=[None], dtype=tf.float32, name="{}_rewards".format(name) |
|
|
|
) |
|
|
|
rewards_holder = tf.placeholder( |
|
|
|
shape=[None], dtype=tf.float32, name="{}_rewards".format(name) |
|
|
|
) |
|
|
|
self.rewards_holders[name] = rewards_holder |
|
|
|
|
|
|
|
q1_losses = [] |
|
|
|