dropout ratio for inputs
copyrightly opened this issue · comments
luwei commented
In the layer.py line 170, there is x = tf.nn.dropout(x, 1-self.dropout)
. By the documentation of tf, the argument for tf.nn.dropout
is the ratio to drop. So should it be x = tf.nn.dropout(x, self.dropout)
?
shannon-np commented
I believe you are correct, but at the time of writing this was not the case.
There was a change in the code for tf.nn.dropout
; according to the documentation,
Warning: SOME ARGUMENTS ARE DEPRECATED: (keep_prob). They will be removed in a future version. Instructions for updating: Please use rate instead of keep_prob. Rate should be set to rate = 1 - keep_prob.