{"id":175,"date":"2020-02-20T20:51:35","date_gmt":"2020-02-20T12:51:35","guid":{"rendered":"http:\/\/10.1.224.228\/?p=175"},"modified":"2022-07-06T11:09:08","modified_gmt":"2022-07-06T03:09:08","slug":"progress-report-clara-medical-imaging","status":"publish","type":"post","link":"https:\/\/novaglobal.com.sg\/index.php\/progress-report-clara-medical-imaging\/","title":{"rendered":"OUR BREAKTHROUGH WORKING WITH NVIDIA CLARA ON BLOOD VESSELS"},"content":{"rendered":"\n<p>After much experimentation, VesselNet\u2019s Keras model has been successfully converted into a Tensorflow-TensorRT optimized model via the following transformation:<\/p>\n\n\n\n<ul><li>A custom function called freeze_session() that freezes the graph\nsession of a trained TF model and saved to a ProtoBuf (.pb) format: <\/li><\/ul>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"724\" height=\"589\" src=\"http:\/\/10.1.224.228\/wp-content\/uploads\/2020\/02\/image.png\" alt=\"\" class=\"wp-image-176\" srcset=\"https:\/\/novaglobal.com.sg\/wp-content\/uploads\/2020\/02\/image.png 724w, https:\/\/novaglobal.com.sg\/wp-content\/uploads\/2020\/02\/image-300x244.png 300w\" sizes=\"(max-width: 724px) 100vw, 724px\" \/><\/figure>\n\n\n\n<ul><li>Utilizing the Python API of TF-TRT, trt.TrtGraphConverter() method to convert the\nprevious ProtoBuf format file into a TRT-Optimized frozen graph:<\/li><\/ul>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"726\" height=\"198\" src=\"http:\/\/10.1.224.228\/wp-content\/uploads\/2020\/02\/image-1.png\" alt=\"\" class=\"wp-image-177\" srcset=\"https:\/\/novaglobal.com.sg\/wp-content\/uploads\/2020\/02\/image-1.png 726w, https:\/\/novaglobal.com.sg\/wp-content\/uploads\/2020\/02\/image-1-300x82.png 300w\" sizes=\"(max-width: 726px) 100vw, 726px\" \/><\/figure>\n\n\n\n<p>The\ncoming week we\u2019ll move towards uploading to AIAA server and run several tests\non it with training data.<\/p>\n\n\n\n<p>Commands to upload to AIAA Server:<\/p>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"726\" height=\"68\" src=\"http:\/\/10.1.224.228\/wp-content\/uploads\/2020\/02\/image-2.png\" alt=\"\" class=\"wp-image-178\" srcset=\"https:\/\/novaglobal.com.sg\/wp-content\/uploads\/2020\/02\/image-2.png 726w, https:\/\/novaglobal.com.sg\/wp-content\/uploads\/2020\/02\/image-2-300x28.png 300w\" sizes=\"(max-width: 726px) 100vw, 726px\" \/><\/figure>\n\n\n\n<p>Here are the\nsamples of the testing datasets provided by the VesselNet\u2019s author:<\/p>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"702\" height=\"375\" src=\"http:\/\/10.1.224.228\/wp-content\/uploads\/2020\/02\/image-3.png\" alt=\"\" class=\"wp-image-179\" srcset=\"https:\/\/novaglobal.com.sg\/wp-content\/uploads\/2020\/02\/image-3.png 702w, https:\/\/novaglobal.com.sg\/wp-content\/uploads\/2020\/02\/image-3-300x160.png 300w\" sizes=\"(max-width: 702px) 100vw, 702px\" \/><\/figure>\n\n\n\n<figure class=\"wp-block-image size-large\"><img loading=\"lazy\" decoding=\"async\" width=\"705\" height=\"403\" src=\"http:\/\/10.1.224.228\/wp-content\/uploads\/2020\/02\/image-4.png\" alt=\"\" class=\"wp-image-180\" srcset=\"https:\/\/novaglobal.com.sg\/wp-content\/uploads\/2020\/02\/image-4.png 705w, https:\/\/novaglobal.com.sg\/wp-content\/uploads\/2020\/02\/image-4-300x171.png 300w\" sizes=\"(max-width: 705px) 100vw, 705px\" \/><\/figure>\n\n\n\n<p><strong>Summary<\/strong><\/p>\n\n\n\n<p>The VesselNet\u2019s Keras model was successfully converted into a TRT-Optimized frozen graph that now can be uploaded to AIAA server to integrate with existing pipeline. Next, we\u2019ll upload and run through the workflow with testing datasets. <\/p>\n","protected":false},"excerpt":{"rendered":"<p>After much experimentation, VesselNet\u2019s Keras model has been successfully converted into a Tensorflow-TensorRT optimized model via the following transformation: A custom function called freeze_session() that freezes the graph session of a trained TF model and saved to a ProtoBuf (.pb) format: Utilizing the Python API of TF-TRT, trt.TrtGraphConverter() method to convert the previous ProtoBuf format [&hellip;]<\/p>\n","protected":false},"author":2,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[],"_links":{"self":[{"href":"https:\/\/novaglobal.com.sg\/index.php\/wp-json\/wp\/v2\/posts\/175"}],"collection":[{"href":"https:\/\/novaglobal.com.sg\/index.php\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/novaglobal.com.sg\/index.php\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/novaglobal.com.sg\/index.php\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/novaglobal.com.sg\/index.php\/wp-json\/wp\/v2\/comments?post=175"}],"version-history":[{"count":3,"href":"https:\/\/novaglobal.com.sg\/index.php\/wp-json\/wp\/v2\/posts\/175\/revisions"}],"predecessor-version":[{"id":184,"href":"https:\/\/novaglobal.com.sg\/index.php\/wp-json\/wp\/v2\/posts\/175\/revisions\/184"}],"wp:attachment":[{"href":"https:\/\/novaglobal.com.sg\/index.php\/wp-json\/wp\/v2\/media?parent=175"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/novaglobal.com.sg\/index.php\/wp-json\/wp\/v2\/categories?post=175"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/novaglobal.com.sg\/index.php\/wp-json\/wp\/v2\/tags?post=175"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}