why did u add this kind of dropout in every residual block

`def layer_dropout(inputs, residual, dropout):
pred = tf.random_uniform([]) < dropout
return tf.cond(pred, lambda: residual, lambda: tf.nn.dropout(inputs, 1.0 - dropout) + residual)`

Start at the layer_dropout entry point shown in the issue and trace where it is used in the residual blocks. Review the repository history and surrounding TensorFlow behavior to determine the rationale for the conditional dropout; done means documenting a supported explanation for this design.

Written by the indexing model from the issue text.