the problem in running about deep_learning HOT 2 OPEN

Hey,

It may indeed come from the change in versions of Keras, the code used to run. However the error seems quite clear : it expected a type of int32 (32 bits integer) and got a Variable with dtype float32 (32bits float).

Can't say for sure, it seems like you've changed the code somehow but I'd bet on this: LSTM() expects an int as number of neurons in the layer but you gave a tensorflow variable from the previous LSTM layer instead. Could simply be that you inverted the order of arguments

from deep_learning.

you're so kind,thank you so much,I found that when I change the backend tensorflow into theano,the code can run smothly.

from deep_learning.