如何在Theano中分配/更新张量共享变量的子集？

Question

When compiling a function in theano , a shared variable(say X) can be updated by specifying updates=[(X, new_value)] . Now I am trying to update only subset of a shared variable:

from theano import tensor as T
from theano import function
import numpy

X = T.shared(numpy.array([0,1,2,3,4]))
Y = T.vector()
f = function([Y], updates=[(X[2:4], Y)] # error occur:
                                        # 'update target must 
                                        # be a SharedVariable'

The codes will raise a error "update target must be a SharedVariable", I guess that means update targets can't be non-shared variables. So is there any way to compile a function to just udpate subset of shared variables?

Answer 1

Use set_subtensor or inc_subtensor :

from theano import tensor as T
from theano import function, shared
import numpy

X = shared(numpy.array([0,1,2,3,4]))
Y = T.vector()
X_update = (X, T.set_subtensor(X[2:4], Y))
f = function([Y], updates=[X_update])
f([100,10])
print X.get_value() # [0 1 100 10 4]

There's now a page about this in the Theano FAQ: http://deeplearning.net/software/theano/tutorial/faq_tutorial.html

Answer 2

This code should solve your problem:

from theano import tensor as T
from theano import function, shared
import numpy

X = shared(numpy.array([0,1,2,3,4], dtype='int'))
Y = T.lvector()
X_update = (X, X[2:4]+Y)
f = function(inputs=[Y], updates=[X_update])
f([100,10])
print X.get_value()
# output: [102 13]

And here is the introduction about shared variables in the official tutorial .

Please ask, if you have further questions!