python - 3d sliding window operation in Theano? -

-

tl.dr. there 3-dimensional friendly implementation of theano.tensor.nnet.neighbours.images2neibs?

i perform voxel-wise classification of volume (nxnxn) using neural network takes in nxnxn image, n>n. classify each voxel in volume, have iterate through each voxel. each iterration, obtain , pass neighborhood voxels input neural network. sliding window operation, operation neural network.

while neural network implemented in theano, sliding window implementation in python/numpy. since not pure theano operation, classification takes forever (> 3 hours) classify voxels in 1 volume. 2d sliding window operation, theano has helper method, theano.tensor.nnet.neighbours.images2neibs, there similar implementation 3-dimensional images?

edit: there existing numpy solutions(1 , 2) n-d sliding window, both use np.lib.stride_tricks.as_strided provide 'views of sliding window', preventing memory issues. in implementation, sliding window arrays being passed numpy (cython) python , theano. boost performance, it's have bypass python.


Comments

Post a Comment

Popular posts from this blog

gridview - Yii2 DataPorivider $totalSum for a column -

-
i have order , ordersearch models. in order list (actionindex) gridview there filtering , sorting. on of columns in order order_total (sum of products in order). i need implement sum of order_total in gridview. if manually counting activedataprovider->getmodels() array_map spend 3 seconds 3000 orders (localhost). don't want miss time. i see 2 ways make faster: create cache every filter , update lifetime (horrible) the interesting can right in ordersearch->search() method $query method. don't understand how can pass in controller. example of code 2nd way: class ordersearch extends order { public $totalsum; public function search($params) { $query = order::find(); $dataprovider = new activedataprovider([ 'query' => $query, ]); $this->load($params); $this->totalsum = $query->sum('order_total'); // works fast return $dataprovider; } } after i
Read more

java - Suppress Jboss version details from HTTP error response -

-
my jboss verson jboss eap 6.3 when kind of error response sent jboss, adds server information response. following sample row response: http/1.1 400 bad request server: apache-coyote/1.1 content-type: text/html;charset=utf-8 content-length: 2391 date: tue, 19 jul 2016 09:30:46 gmt connection: close jboss web/7.4.8.final-redhat-4 - jbweb000064: error report jbweb000065: http status 400 - org.codehaus.jackson.map.jsonmappingexception: can not construct instance of mybean.type string value 'xyz': value not 1 of declared enum instance names @ [source: org.jboss.resteasy.core.interception.messagebodyreadercontextimpl$inputstreamwrapper@5e421c2b; line: 14, column: 26] (through reference chain: requestargs["request"]) jbweb000309: type jbweb000067: status report jbweb000068: message org.codehaus.jackson.map.jsonmappingexception: can not construct instance of mybean.type string value 'xyz': value not 1 of declared enum instance names @ [source: org.jboss
Read more

Sass watch command compiles .scss files before full sftp upload -

-
i use winscp download , edit .scss files , sass on linux (on server) compile them .css . after saving file, use: sass scss/style.scss css/style.css , replaces css file compiled sccs. the problem i want skip part return command line after editing scss file, sake of automation , saving time. but, if use watch command: sass --watch scss/style.scss:css/style.css synchronize 2 files, alerts nonexistent css errors pop up: change detected to: scss/style.scss error scss/style.scss (line 232: invalid css after "...ht: bold; line-": expected "{", "") note that when file uploads quicker usual (sometimes happens) watch command job, no errors. this because, when file upload takes time, sass compiles scss file (on server) before uploaded remote folder. thus, compiles part of file, resulting in css errors. is there way set timer watch command waits few seconds after detects changes , before compiling? any other way overcome is, of course,
Read more