add axis=1 to argmax call in logreg docs to fit new Theano behavior

fbreuer · fbreuer · commit 940cb40f9adf · 2013-01-10T23:24:11.000-08:00
diff --git a/doc/logreg.txt b/doc/logreg.txt
@@ -83,7 +83,7 @@ The code to do this in Theano is the following:
 
   # symbolic description of how to compute prediction as class whose probability
   # is maximal
-  y_pred = T.argmax(p_y_given_x)
+  y_pred = T.argmax(p_y_given_x, axis=1)
 
   # compiled theano function that returns this value
   classify = theano.function(inputs=[x], outputs=y_pred)
