when do use borrow=True for theano shared variables?

Question

In theano when do we use borrow =true for shared variables?

In theano deep learning tutorial, I found that in some examples borrow=True is used while in others it isn't. In the logistic regression and the MLP tutorial, borrow=True is used, while in the Convolutional neural network tutorial borrow=True is not used.

Answer 1

Found answer at theano google group https://groups.google.com/forum/#!topic/theano-users/XWDOoKdpn5I