{"id":24399,"date":"2021-01-17T16:26:44","date_gmt":"2021-01-17T14:26:44","guid":{"rendered":"https:\/\/hgpu.org\/?p=24399"},"modified":"2021-01-17T16:26:44","modified_gmt":"2021-01-17T14:26:44","slug":"fast-convolutional-neural-networks-on-fpgas-with-hls4ml","status":"publish","type":"post","link":"https:\/\/hgpu.org\/?p=24399","title":{"rendered":"Fast convolutional neural networks on FPGAs with hls4ml"},"content":{"rendered":"<p>We introduce an automated tool for deploying ultra low-latency, low-power deep neural networks with large convolutional layers on FPGAs. By extending the hls4ml library, we demonstrate how to achieve inference latency of 5\u03bcs using convolutional architectures, while preserving state-of-the-art model performance. Considering benchmark models trained on the Street View House Numbers Dataset, we demonstrate various methods for model compression in order to fit the computational constraints of a typical FPGA device. In particular, we discuss pruning and quantization-aware training, and demonstrate how resource utilization can be reduced by over 90% while maintaining the original model accuracy.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>We introduce an automated tool for deploying ultra low-latency, low-power deep neural networks with large convolutional layers on FPGAs. By extending the hls4ml library, we demonstrate how to achieve inference latency of 5\u03bcs using convolutional architectures, while preserving state-of-the-art model performance. Considering benchmark models trained on the Street View House Numbers Dataset, we demonstrate various [&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":[11,3],"tags":[451,1782,1673,377,34,176],"class_list":["post-24399","post","type-post","status-publish","format-standard","hentry","category-computer-science","category-paper","tag-benchmarking","tag-computer-science","tag-deep-learning","tag-fpga","tag-neural-networks","tag-package"],"views":1890,"jetpack_publicize_connections":[],"jetpack_featured_media_url":"","jetpack_sharing_enabled":true,"_links":{"self":[{"href":"https:\/\/hgpu.org\/index.php?rest_route=\/wp\/v2\/posts\/24399","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=24399"}],"version-history":[{"count":0,"href":"https:\/\/hgpu.org\/index.php?rest_route=\/wp\/v2\/posts\/24399\/revisions"}],"wp:attachment":[{"href":"https:\/\/hgpu.org\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=24399"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/hgpu.org\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=24399"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/hgpu.org\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=24399"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}