Hi, borrow=True is an indication that the internal value _can_ be returned, but it is not a guarantee that it _will_ be. Therefore, the better way is to explicitly re-assign the value calling set_value, which is guaranteed to work.
On Thu, Oct 27, 2016, pietro galliani wrote: > Apologies if I am misunderstanding something or this is not the correct > venue, but I believe I found a small mistake in the deeplearning.net theano > tutorial. > > In the "Seeding Streams" section of the "More Examples" section, it is > written the following: > > Random variables can be seeded individually or collectively. > > You can seed just one random variable by seeding or assigning to the .rng > attribute, > using .rng.set_value(). > > >>> rng_val = rv_u.rng.get_value(borrow=True) # Get the rng for rv_u>>> > >>> rng_val.seed(89234) # seeds the generator>>> > >>> rv_u.rng.set_value(rng_val, borrow=True) # Assign back seeded rng > > > > Now, it seems to me that using rng.set_value is entirely superfluous here: > since we used borrow=True with get_value, rng_val is the random state of > the (uniform, in this case) random function rv_u and NOT a copy, and, thus, > there is no need to assigning it back to it after seeding it. This would be > different, of course, if we had used borrow=False. > > A couple of quick tests seem to confirm my impression. Am I missing > something? > > Thanks! > > -- > > --- > You received this message because you are subscribed to the Google Groups > "theano-users" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to theano-users+unsubscr...@googlegroups.com. > For more options, visit https://groups.google.com/d/optout. -- Pascal -- --- You received this message because you are subscribed to the Google Groups "theano-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to theano-users+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
