16a31e2c9fedc654e9117b42b8144adf1d0e4900,examples/reinforcement_learning/tutorial_A3C.py,Worker,work,#Worker#Any#,187

Before Change

                    buffer_v_target.reverse()

                    buffer_s, buffer_a, buffer_v_target = (
                        np.vstack(buffer_s), np.vstack(buffer_a), np.vstack(buffer_v_target)
                    )
                    // update gradients on global network
                    self.AC.update_global(buffer_s, buffer_a, buffer_v_target.astype("float32"), globalAC)
                    buffer_s, buffer_a, buffer_r = [], [], []

After Change

                    buffer_s = tf.convert_to_tensor(np.vstack(buffer_s))
                    buffer_a = tf.convert_to_tensor(np.vstack(buffer_a))
                    buffer_v_target = tf.convert_to_tensor(np.vstack(buffer_v_target).astype("float32"))

                    // update gradients on global network
                    self.AC.update_global(buffer_s, buffer_a, buffer_v_target, globalAC)

Instances