The problem of using pylearn2's cnn

[YangShuguang]

firstly,when I configure the pool_style: 'mean'. the computer raise the problem that "ValueError: monitor channels are expected to have dtype float64 but "test_h0_range_x_min_u" has dtype float32"
then,In the pylearn2.models.mlp.ConvRectifiedLinear.mean_pool() function, I change " wide_infinity = T.alloc(-np.inf,........)" to "wide_infinity = T.alloc(T.constant(-np.inf, dtype=config.floatX),.....)". the problem seems to be solved. but the new proble appears, the computer raise the problem that "Exception: NaN in h0_W" .

Am I missing anything?

[YangShuguang]

I want to know that if I want to use 'pool_style: 'mean' ' in the cnn of the pylearn2, what shold I do? Ladies and gentlemen, can you give me an answer or the example of the 'pool_style=mean'? thank you