{"id":1406211,"date":"2023-12-12T17:00:13","date_gmt":"2023-12-12T09:00:13","guid":{"rendered":"https:\/\/makerpro.cc\/makerlearn\/?post_type=courses&p=1406211"},"modified":"2024-01-01T21:50:11","modified_gmt":"2024-01-01T13:50:11","slug":"0907yolov8-tensorrt","status":"publish","type":"courses","link":"https:\/\/makerpro.cc\/makerlearn\/courses\/0907yolov8-tensorrt\/","title":{"rendered":"\u301010\u500d\u901fAI\u904b\u7b97\u3011WSL YOLOv8\u8f49TensorRT\u79d8\u6280\u516c\u958b"},"content":{"rendered":"

YOLOv8\u662f\u6700\u65b0\u7248\u672c\u7684YOLO\u6f14\u7b97\u6cd5\uff0c\u6548\u80fd\u6bd4\u50b3\u7d71darknet\u6846\u67b6\u5feb\u4e4b\u5916\uff0c\u5b89\u88dd\u8207\u652f\u63f4\u4e5f\u90fd\u66f4\u52a0\u7c21\u5316\u53ca\u5b8c\u6574\uff0c\u9019\u6b21\u6211\u5011\u8981\u4ecb\u7d39YOLOv8\u5728NVIDIA\u986f\u793a\u5361\u7684GPU\u52a0\u901f\u5916\uff0c\u66f4\u80fd\u900f\u904eTensorRT\u7684\u67b6\u69cb\u4e0b\u9032\u4e00\u6b65\u63d0\u6607\u6548\u80fd\uff0c\u672c\u6b21\u5be6\u505a\u5c07\u63a1\u7528WSL\u74b0\u5883\uff0c\u8b93\u4f60\u5728Windows\u4e0b\u5b8c\u6210Linux\u7684TensorRT\u6240\u5e36\u4f86\u7684\u6548\u80fd\u63d0\u6607\u3002<\/p>\n

\u3010\u8ab2\u7a0b\u55ae\u5143\u3011<\/h3>\n

\"\"CPU VS GPU<\/h3>\n

\u591c\u5e02\u5c0f\u9738\u738b\u8209\u6b77\u4ee3\u300e\u53e4\u5893\u5947\u5175\u300f\u96fb\u73a9\u4e3b\u89d2\u863f\u62c9\uff0c\u8aaa\u660e\u986f\u793a\u5361\u76843D\u904b\u7b97\u529f\u80fd\uff0c\u96a8\u8457\u986f\u793a\u5361\u7684\u6539\u671d\u63db\u4ee3\uff0c\u904a\u6232\u4e3b\u89d2\u7684\u5916\u578b\u8d8a\u898b\u7d30\u7dfb\uff0cGPU\u70ba\u4e86\u8981\u5f37\u5316\u904a\u6232\u6548\u679c\uff0c\u5c31\u52a0\u5165\u4e86\u5f88\u591a\u6838\u5fc3\u53bb\u52a0\u901f\u904b\u7b97\u6548\u80fd\u3002\u5f8c\u4f86\u56e0\u70ba3D\u904b\u7b97\u9700\u6c42\u63d0\u9ad8(AI\u3001\u6316\u7926)\uff0c\u5c31\u6709\u4eba\u958b\u59cb\u6e2c\u8a66\u9ad8\u968e\u986f\u793a\u5361\u8207\u7684\u904b\u7b97\u6548\u80fd\u3002\u5f97\u5230\u4ee4\u4eba\u6eff\u610f\u7684\u6548\u679c\uff0c\u5f8c\u4f86\u5728\u8655\u7406AI\u904b\u7b97\u6642\u5c31\u504f\u91cdGPU\u7684\u4f7f\u7528\u3002<\/p>\n

\"\"<\/p>\n

\"\"<\/p>\n

\u4ec0\u9ebc\u662f CUDA?<\/h3>\n

NVIDIA\u63a8\u51fa\u7684Compute Unified Device Architecture
\n\u4e3b\u8981\u4f5c\u7528\u662f\u9023\u63a5GPU \u548c \u61c9\u7528\u7a0b\u5f0f\uff0c\u65b9\u4fbf\u4f7f\u7528\u8005\u901a\u904e CUDA \u7684 API \u8abf\u5ea6GPU\u7684core\u9032\u884c\u8a08\u7b97\u3002<\/p>\n

\"\"<\/p>\n

\u4ec0\u9ebc\u662fcuDNN<\/h3>\n

nVIDIA\u63a8\u51fa\u7684cuDNN\uff1a<\/p>\n

cuDA Deep Neural Network library
\n\u6df1\u5ea6\u795e\u7d93\u7db2\u8def(DNN)\u7684\u52a0\u901f\u7a0b\u5f0f\u5eab\uff0c\u5c07\u6a21\u578b\u8a13\u7df4\u7684\u8a08\u7b97\u512a\u5316\u4e4b\u5f8c\uff0c\u518d\u901a\u904e CUDA \u8abf\u5ea6GPU\u7684 core \u9032\u884c<\/p>\n

https:\/\/developer.nvidia.com\/cuda-gpus<\/p>\n

\"\"<\/p>\n

 <\/p>\n

\u4ec0\u9ebc\u662fTensorRT<\/h3>\n

\u53ea\u8ca0\u8cac\u6a21\u578b\u7684\u63a8\u7406\uff08inference\uff09\u904e\u7a0b\uff0cTensorRT\u4e0d\u7528\u65bc\u8a13\u7df4\u6a21\u578b\u7684\uff0c\u800c\u662f\u7528\u65bc\u90e8\u7f72\u6642\u52a0\u901f\u6a21\u578b\u904b\u884c\u901f\u5ea6\u3002\u6bd4\u5982\u7528\u5728\u8eca\u6d41\/\u4eba\u6d41\u8fa8\u8b58\u904b\u7b97\u7b49\u9700\u8981\u9ad8\u901f\u5f71\u50cf\u8655\u7406\uff0c\u5c31\u6703\u7d50\u5408\u9019\u500b\u6846\u67b6\u3002<\/p>\n

\"\"<\/p>\n

 <\/p>\n

\"\"<\/p>\n

\"\"<\/p>\n

\u591c\u5e02\u5c0f\u9738\u738b\u5e36\u9818\u5148\u5b89\u88ddWSL\u8207MS\u7cfb\u7d71\u4e2d\u5b89\u88ddVSCode\u3002\u5b89\u88dd\u9060\u7aef\u958b\u767c\u5de5\u5177\u3001\u555f\u52d5Win\u7aefVS Code\u3002\u53e6\u5916\u4e5f\u63d0\u4f9bWSL\u8b1b\u7fa9\u7d66\u521d\u6b21\u63a5\u89f8WSL\u7684\u5b78\u54e1\u4e0b\u8f09\u95b1\u8b80\u3002<\/p>\n

\"\"<\/p>\n

\"\"<\/p>\n

\"\"<\/p>\n

YOLOv8\u4f7f\u7528 PyTorch\u958b\u767c\uff0c\u5df2\u652f\u63f4\u76ee\u6a19\u6aa2\u6e2c\u8207\u8ddf\u8e64\u3001\u5be6\u4f8b\u5206\u5272\u3001\u5716\u50cf\u5206\u985e\u548c\u59ff\u614b\u4f30\u8a08\u7b49\u4efb\u52d9\u3002\u672c\u55ae\u5143\u5c07\u6559\u60a8\u5982\u4f55\u5728WSL\u4e0a\u5b89\u88dd\u8a2d\u5b9aYOLOv8\u3001\u8a13\u7df4\u81ea\u5df1\u7684\u8cc7\u6599\u96c6\uff0c\u4e26\u6e2c\u8a66\u8a13\u7df4\u51fa\u7684\u7db2\u8def\u6a21\u578b\u3002<\/p>\n

\"\"<\/p>\n

TensorRT \u662fNVIDIA \u91dd\u5c0d\u81ea\u5df1\u7684\u7522\u54c1\u6240\u505a\u7684\u4e00\u500b\u5de5\u5177\uff0c\u4e3b\u8981\u76ee\u7684\u5728\u65bc\u5354\u52a9\u4f7f\u7528\u8005\u7684\u6a21\u578b\uff0c\u80fd\u66f4\u5feb\u66f4\u6709\u6548\u7684\u53d6\u7528\u81ea\u5bb6GPU \u8cc7\u6e90\uff0c\u4e26\u900f\u904e\u4e00\u4e9b\u512a\u5316\u7a0b\u5e8f\uff0c\u52a0\u901f\u6a21\u578b\u7684\u63a8\u8ad6\u3002 TensorRT\u53ef\u4ee5\u5c07YOLO\u3001TensorFlow\u3001PyTorch\u548cCaffe\u7b49AI\u6846\u67b6\u8a13\u7df4\u597d\u7684\u6a21\u578b\u8f49\u63db\u6210\u9ad8\u5ea6\u512a\u5316\u7684\u63a8\u8ad6\u5f15\u64ce\uff0c\u518d\u90e8\u7f72\u5230NVIDIA\u7684\u5404\u985e\u5e73\u53f0\u4e0a\u9032\u884c\u5373\u6642\u7684\u63a8\u8ad6\u61c9\u7528\u3002\u672c\u55ae\u5143\u5c07\u6559\u60a8\u5982\u4f55\u7528TensorRT\u4f86\u8f49\u63dbYOLOv8\u6a21\u578b\u3002<\/p>\n

\"\"<\/p>\n

\u672c\u55ae\u5143\u5c07\u4ee5\u5be6\u969b\u500b\u6848\u4f86\u8aaa\u660e\u7528TensorRT\u4f86\u8f49\u63dbYOLOv8\u6a21\u578b\u7684\u904b\u884c\u6548\u76ca\u3002<\/p>\n","protected":false},"excerpt":{"rendered":"

YOLOv8\u662f\u6700\u65b0\u7248\u672c\u7684YOLO\u6f14\u7b97\u6cd5\uff0c\u6548\u80fd\u6bd4\u50b3\u7d71darknet\u6846\u67b6\u5feb\u4e4b\u5916\uff0c\u5b89\u88dd\u8207\u652f\u63f4\u4e5f\u90fd\u66f4\u52a0\u7c21\u5316\u53ca\u5b8c\u6574\uff0c\u9019\u6b21\u6211\u5011\u8981\u4ecb\u7d39YOLOv8\u5728NVIDIA\u986f\u793a\u5361\u7684GPU\u52a0\u901f\u5916\uff0c\u66f4\u80fd\u900f\u904eTensorRT\u7684\u67b6\u69cb\u4e0b\u9032\u4e00\u6b65\u63d0\u6607\u6548\u80fd\uff0c\u672c\u6b21\u5be6\u505a\u5c07\u63a1\u7528WSL\u74b0\u5883\uff0c\u8b93\u4f60\u5728Windows\u4e0b\u5b8c\u6210Linux\u7684TensorRT\u6240\u5e36\u4f86\u7684\u6548\u80fd\u63d0\u6607\u3002<\/p>\n","protected":false},"author":9883,"featured_media":1409116,"template":"","course-category":[84,88,166,93],"course-tag":[233,825,800,870,871,869],"class_list":["post-1406211","courses","type-courses","status-publish","has-post-thumbnail","hentry","course-category-edge-ai","course-category-ai","course-category-course","course-category-category","course-tag-linux","course-tag-nvidia","course-tag-tensorflow","course-tag-tensorrt","course-tag-windows","course-tag-yolov8","description-off"],"acf":[],"jetpack_sharing_enabled":true,"_links":{"self":[{"href":"https:\/\/makerpro.cc\/makerlearn\/wp-json\/wp\/v2\/courses\/1406211","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/makerpro.cc\/makerlearn\/wp-json\/wp\/v2\/courses"}],"about":[{"href":"https:\/\/makerpro.cc\/makerlearn\/wp-json\/wp\/v2\/types\/courses"}],"author":[{"embeddable":true,"href":"https:\/\/makerpro.cc\/makerlearn\/wp-json\/wp\/v2\/users\/9883"}],"version-history":[{"count":10,"href":"https:\/\/makerpro.cc\/makerlearn\/wp-json\/wp\/v2\/courses\/1406211\/revisions"}],"predecessor-version":[{"id":1409118,"href":"https:\/\/makerpro.cc\/makerlearn\/wp-json\/wp\/v2\/courses\/1406211\/revisions\/1409118"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/makerpro.cc\/makerlearn\/wp-json\/wp\/v2\/media\/1409116"}],"wp:attachment":[{"href":"https:\/\/makerpro.cc\/makerlearn\/wp-json\/wp\/v2\/media?parent=1406211"}],"wp:term":[{"taxonomy":"course-category","embeddable":true,"href":"https:\/\/makerpro.cc\/makerlearn\/wp-json\/wp\/v2\/course-category?post=1406211"},{"taxonomy":"course-tag","embeddable":true,"href":"https:\/\/makerpro.cc\/makerlearn\/wp-json\/wp\/v2\/course-tag?post=1406211"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}