About dropout

[pitera05]

Hi @ line 103, may I know if it's typo? since the keep_prob will be 0 when dropout is 1, it's strange.

      if dropout > 0.0:
        output = tf.nn.dropout(output, keep_prob=1 - dropout)