python - Theano assertion error -


when run theano , want train 2d convolutional network following error:

error (theano.gof.opt): optimization failure due to: local_abstractconv_check error (theano.gof.opt): node: abstractconv2d{border_mode='half', subsample=(1, 1), filter_flip=true, imshp=(none, none, none, none), kshp=(32, 3, 3, 3), filter_dilation=(1, 1)}(convolution2d_input_1, convolution2d_1_w) error (theano.gof.opt): traceback: error (theano.gof.opt): traceback (most recent call last):   file "c:\users\nicol\anaconda3\envs\py34\lib\site-packages\theano\gof\opt.py", line 1820, in process_node     replacements = lopt.transform(node)   file "c:\users\nicol\anaconda3\envs\py34\lib\site-packages\theano\tensor\nnet\opt.py", line 409, in local_abstractconv_check     node.op.__class__.__name__) assertionerror: abstractconv2d theano optimization failed: there no implementation available supporting requested options. did exclude both "conv_dnn" , "conv_gemm" optimizer? if on gpu, cudnn available , gpu support it? if on cpu, have blas library installed theano can link against? 

i have anaconda (which includes mkl) installed python 3.4 latest version of theano. installed mingw , libpython. checked presence of blas running check_blas.py script provided theanom said theano connected blas through numpy.

i not wrong , still (obiously) want optimizers work. have seen several posts online problem none of them got real solution other installing else or not using optimizers. know how prevent error?


Comments

Popular posts from this blog

java - Jasper subreport showing only one entry from the JSON data source when embedded in the Title band -

serialization - Convert Any type in scala to Array[Byte] and back -

SonarQube Plugin for Jenkins does not find SonarQube Scanner executable -