{"id":12965,"date":"2014-10-22T20:54:02","date_gmt":"2014-10-22T17:54:02","guid":{"rendered":"http:\/\/hgpu.org\/?p=12965"},"modified":"2014-10-22T20:54:02","modified_gmt":"2014-10-22T17:54:02","slug":"introducing-currennt-the-munich-open-source-cuda-recurrent-neural-network-toolkit","status":"publish","type":"post","link":"https:\/\/hgpu.org\/?p=12965","title":{"rendered":"Introducing CURRENNT &#8211; the Munich open-source CUDA RecurREnt Neural Network Toolkit"},"content":{"rendered":"<p>In this article, we introduce CURRENNT, an open-source parallel implementation of deep recurrent neural networks (RNNs) supporting graphics processing units (GPUs) through NVIDIA&#8217;s Computed Unified Device Architecture (CUDA). CURRENNT supports uni- and bidirectional RNNs with Long Short-Term Memory (LSTM) memory cells which overcome the vanishing gradient problem. To our knowledge, CURRENNT is the first publicly available parallel implementation of deep LSTM-RNNs. Benchmarks are given on a noisy speech recognition task from the 2013 2nd CHiME Speech Separation and Recognition Challenge, where LSTM-RNNs have been shown to deliver best performance. In the result, double digit speedups in bidirectional LSTM training are achieved with respect to a reference single-threaded CPU implementation. CURRENNT is available under the GNU General Public License from http:\/\/sourceforge.net\/p\/currennt.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>In this article, we introduce CURRENNT, an open-source parallel implementation of deep recurrent neural networks (RNNs) supporting graphics processing units (GPUs) through NVIDIA&#8217;s Computed Unified Device Architecture (CUDA). CURRENNT supports uni- and bidirectional RNNs with Long Short-Term Memory (LSTM) memory cells which overcome the vanishing gradient problem. To our knowledge, CURRENNT is the first publicly [&hellip;]<\/p>\n","protected":false},"author":351,"featured_media":0,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"_jetpack_memberships_contains_paid_content":false,"footnotes":"","jetpack_publicize_message":"","jetpack_publicize_feature_enabled":true,"jetpack_social_post_already_shared":true,"jetpack_social_options":{"image_generator_settings":{"template":"highway","default_image_id":0,"font":"","enabled":false},"version":2}},"categories":[11,89,3],"tags":[1782,14,34,20,1090,176,848],"class_list":["post-12965","post","type-post","status-publish","format-standard","hentry","category-computer-science","category-nvidia-cuda","category-paper","tag-computer-science","tag-cuda","tag-neural-networks","tag-nvidia","tag-nvidia-geforce-gtx-560","tag-package","tag-speech-recognition"],"views":4647,"jetpack_publicize_connections":[],"jetpack_featured_media_url":"","jetpack_sharing_enabled":true,"_links":{"self":[{"href":"https:\/\/hgpu.org\/index.php?rest_route=\/wp\/v2\/posts\/12965","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/hgpu.org\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/hgpu.org\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/hgpu.org\/index.php?rest_route=\/wp\/v2\/users\/351"}],"replies":[{"embeddable":true,"href":"https:\/\/hgpu.org\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=12965"}],"version-history":[{"count":0,"href":"https:\/\/hgpu.org\/index.php?rest_route=\/wp\/v2\/posts\/12965\/revisions"}],"wp:attachment":[{"href":"https:\/\/hgpu.org\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=12965"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/hgpu.org\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=12965"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/hgpu.org\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=12965"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}