{"id":477761,"date":"2023-08-09T09:19:52","date_gmt":"2023-08-09T09:19:52","guid":{"rendered":""},"modified":"2023-09-05T11:15:22","modified_gmt":"2023-09-05T11:15:22","slug":"keras","status":"publish","type":"wiki","link":"https:\/\/oneproxy.pro\/cn\/wiki\/keras\/","title":{"rendered":"\u5580\u62c9\u65af"},"content":{"rendered":"<p>Keras \u662f\u4e00\u4e2a\u7528 Python \u7f16\u5199\u7684\u5f00\u6e90\u6df1\u5ea6\u5b66\u4e60\u6846\u67b6\u3002\u5b83\u4ee5\u5176\u7528\u6237\u53cb\u597d\u6027\u548c\u7075\u6d3b\u6027\u800c\u95fb\u540d\uff0c\u4f7f\u5176\u6210\u4e3a\u7814\u7a76\u4eba\u5458\u3001\u5f00\u53d1\u4eba\u5458\u548c\u6570\u636e\u79d1\u5b66\u5bb6\u6784\u5efa\u548c\u5b9e\u9a8c\u795e\u7ecf\u7f51\u7edc\u7684\u70ed\u95e8\u9009\u62e9\u3002 Keras \u6700\u521d\u7531 Fran\u00e7ois Chollet \u4e8e 2015 \u5e74\u4f5c\u4e3a\u72ec\u7acb\u9879\u76ee\u5f00\u53d1\uff0c\u540e\u6765\u96c6\u6210\u5230 TensorFlow \u5e93\u4e2d\uff0c\u6210\u4e3a\u5176\u5b98\u65b9\u9ad8\u7ea7 API\u3002\u8be5\u6846\u67b6\u5141\u8bb8\u7528\u6237\u4ee5\u6700\u5c0f\u7684\u52aa\u529b\u5b9a\u4e49\u548c\u8bad\u7ec3\u590d\u6742\u7684\u795e\u7ecf\u7f51\u7edc\u6a21\u578b\uff0c\u4f7f\u6df1\u5ea6\u5b66\u4e60\u9886\u57df\u7684\u521d\u5b66\u8005\u548c\u4e13\u5bb6\u90fd\u53ef\u4ee5\u4f7f\u7528\u5b83\u3002<\/p>\n<h2>Keras \u7684\u8d77\u6e90\u5386\u53f2\u548c\u9996\u6b21\u63d0\u53ca<\/h2>\n<p>Keras \u7684\u5386\u53f2\u53ef\u4ee5\u8ffd\u6eaf\u5230 2010 \u5e74\u4ee3\u521d\uff0c\u5f53\u65f6 Fran\u00e7ois Chollet \u5f00\u59cb\u4ee5\u4e2a\u4eba\u540d\u4e49\u81f4\u529b\u4e8e\u8be5\u9879\u76ee\u3002\u4ed6\u7684\u4e3b\u8981\u76ee\u6807\u662f\u521b\u5efa\u4e00\u4e2a\u7528\u6237\u53cb\u597d\u7684\u6df1\u5ea6\u5b66\u4e60\u6846\u67b6\uff0c\u4ee5\u5b9e\u73b0\u5feb\u901f\u5b9e\u9a8c\u548c\u539f\u578b\u8bbe\u8ba1\u3002 2015\u5e743\u6708\uff0cFran\u00e7ois\u5728GitHub\u4e0a\u6b63\u5f0f\u53d1\u5e03Keras\uff0c\u8fc5\u901f\u83b7\u5f97\u6df1\u5ea6\u5b66\u4e60\u793e\u533a\u7684\u8ba4\u53ef\u548c\u8d5e\u8d4f\u3002<\/p>\n<p>Keras \u56e0\u5176\u72ec\u7279\u7684\u8bbe\u8ba1\uff08\u4f18\u5148\u8003\u8651\u7b80\u5355\u6027\u548c\u6613\u7528\u6027\uff09\u800c\u9996\u6b21\u53d7\u5230\u5173\u6ce8\u3002\u6df1\u5ea6\u5b66\u4e60\u7231\u597d\u8005\u548c\u7814\u7a76\u4eba\u5458\u5bf9\u5176\u76f4\u89c2\u7684 API \u5f88\u611f\u5174\u8da3\uff0c\u8fd9\u4f7f\u4ed6\u4eec\u80fd\u591f\u4e13\u6ce8\u4e8e\u6784\u5efa\u521b\u65b0\u6a21\u578b\uff0c\u800c\u4e0d\u4f1a\u8ff7\u5931\u5728\u4f4e\u7ea7\u7ec6\u8282\u7684\u590d\u6742\u6027\u4e2d\u3002<\/p>\n<h2>\u6709\u5173 Keras \u7684\u8be6\u7ec6\u4fe1\u606f\u3002\u6269\u5c55 Keras \u4e3b\u9898<\/h2>\n<p>Keras \u5efa\u7acb\u5728\u6a21\u5757\u5316\u548c\u53ef\u6269\u5c55\u6027\u7684\u539f\u5219\u4e4b\u4e0a\u3002\u5b83\u63d0\u4f9b\u4e86\u5e7f\u6cdb\u7684\u9884\u6784\u5efa\u5c42\u3001\u6fc0\u6d3b\u51fd\u6570\u3001\u4f18\u5316\u7b97\u6cd5\u548c\u635f\u5931\u51fd\u6570\u3002\u8fd9\u79cd\u6a21\u5757\u5316\u65b9\u6cd5\u901a\u8fc7\u5806\u53e0\u6216\u8fde\u63a5\u8fd9\u4e9b\u9884\u5b9a\u4e49\u7684\u7ec4\u4ef6\u6765\u4fc3\u8fdb\u590d\u6742\u795e\u7ecf\u7f51\u7edc\u7684\u6784\u5efa\u3002\u6b64\u5916\uff0cKeras \u901a\u8fc7\u5176\u529f\u80fd API \u63d0\u4f9b\u5b9a\u5236\u6a21\u578b\u4ee5\u6ee1\u8db3\u7279\u5b9a\u8981\u6c42\u7684\u81ea\u7531\uff0c\u4ece\u800c\u5141\u8bb8\u591a\u8f93\u5165\u548c\u591a\u8f93\u51fa\u67b6\u6784\u3002<\/p>\n<p>\u901a\u8fc7\u4e0e TensorFlow \u65e0\u7f1d\u96c6\u6210\uff0cKeras \u53d7\u76ca\u4e8e TensorFlow \u7684\u9ad8\u7ea7\u529f\u80fd\u3001\u53ef\u6269\u5c55\u6027\u548c\u90e8\u7f72\u9009\u9879\u3002\u8fd9\u79cd\u96c6\u6210\u4e3a Keras \u5728\u751f\u4ea7\u7ea7\u5e94\u7528\u7a0b\u5e8f\u548c\u5927\u89c4\u6a21\u6df1\u5ea6\u5b66\u4e60\u9879\u76ee\u4e2d\u4f7f\u7528\u63d0\u4f9b\u4e86\u673a\u4f1a\u3002<\/p>\n<h2>Keras\u7684\u5185\u90e8\u7ed3\u6784\u3002 Keras \u7684\u5de5\u4f5c\u539f\u7406<\/h2>\n<p>Keras \u9075\u5faa\u9ad8\u7ea7 API \u8bbe\u8ba1\uff0c\u62bd\u8c61\u4e86\u6df1\u5ea6\u5b66\u4e60\u7684\u590d\u6742\u6027\u3002\u5176\u67b6\u6784\u5206\u4e3a\u4e09\u4e2a\u4e3b\u8981\u7ec4\u4ef6\uff1a<\/p>\n<ol>\n<li>\n<p><strong>\u540e\u7aef\uff1a<\/strong> \u540e\u7aef\u8d1f\u8d23\u6267\u884cKeras\u7684\u8ba1\u7b97\u64cd\u4f5c\u3002\u7528\u6237\u53ef\u4ee5\u6839\u636e\u81ea\u5df1\u7684\u504f\u597d\u6216\u786c\u4ef6\u517c\u5bb9\u6027\u7075\u6d3b\u9009\u62e9\u4e0d\u540c\u7684\u540e\u7aef\u5f15\u64ce\uff0c\u4f8b\u5982 TensorFlow\u3001Theano \u6216 CNTK\u3002<\/p>\n<\/li>\n<li>\n<p><strong>\u5c42\u6570\uff1a<\/strong> Keras \u63d0\u4f9b\u4e86\u591a\u79cd\u5c42\uff0c\u5305\u62ec\u5bc6\u96c6\uff08\u5168\u8fde\u63a5\uff09\u5c42\u3001\u5377\u79ef\u5c42\u3001\u5faa\u73af\u5c42\u3001\u6c60\u5316\u5c42\u7b49\u3002\u8fd9\u4e9b\u5c42\u53ef\u4ee5\u7ec4\u5408\u548c\u5806\u53e0\u4ee5\u521b\u5efa\u5f3a\u5927\u7684\u795e\u7ecf\u7f51\u7edc\u67b6\u6784\u3002<\/p>\n<\/li>\n<li>\n<p><strong>\u6977\u6a21\uff1a<\/strong> Keras \u7684\u6838\u5fc3\u6784\u5efa\u5757\u662f Model \u7c7b\uff0c\u5b83\u5141\u8bb8\u7528\u6237\u5c06\u5c42\u7ec4\u7ec7\u6210\u5185\u805a\u7ed3\u6784\u4ee5\u5f62\u6210\u795e\u7ecf\u7f51\u7edc\u3002 Keras \u652f\u6301\u9002\u7528\u4e8e\u7ebf\u6027\u5806\u6808\u5f0f\u67b6\u6784\u7684\u987a\u5e8f\u6a21\u578b\uff0c\u4ee5\u53ca\u9002\u7528\u4e8e\u66f4\u590d\u6742\u7684\u591a\u8f93\u5165\u548c\u591a\u8f93\u51fa\u7f51\u7edc\u7684\u529f\u80fd API\u3002<\/p>\n<\/li>\n<\/ol>\n<h2>Keras\u5173\u952e\u7279\u6027\u5206\u6790<\/h2>\n<p>Keras \u56e0\u5176\u9c9c\u660e\u7684\u7279\u70b9\u800c\u5728\u6df1\u5ea6\u5b66\u4e60\u6846\u67b6\u4e2d\u8131\u9896\u800c\u51fa\uff1a<\/p>\n<ol>\n<li>\n<p><strong>\u65b9\u4fbf\u4f7f\u7528\u7684\uff1a<\/strong> Keras \u63d0\u4f9b\u4e86\u76f4\u89c2\u4e14\u7b80\u5355\u7684 API\uff0c\u4f7f\u521d\u5b66\u8005\u53ef\u4ee5\u8f7b\u677e\u4e0a\u624b\u6df1\u5ea6\u5b66\u4e60\u3002<\/p>\n<\/li>\n<li>\n<p><strong>\u6a21\u5757\u5316\uff1a<\/strong> \u8be5\u6846\u67b6\u7684\u6a21\u5757\u5316\u8bbe\u8ba1\u5141\u8bb8\u7528\u6237\u901a\u8fc7\u7ec4\u5408\u9884\u6784\u5efa\u7684\u7ec4\u4ef6\u6765\u6784\u5efa\u548c\u8bd5\u9a8c\u795e\u7ecf\u7f51\u7edc\u3002<\/p>\n<\/li>\n<li>\n<p><strong>\u7075\u6d3b\u6027\uff1a<\/strong> \u51ed\u501f\u591a\u4e2a\u540e\u7aef\u9009\u9879\u4ee5\u53ca\u4e0e TensorFlow \u7684\u65e0\u7f1d\u96c6\u6210\uff0cKeras \u53ef\u4ee5\u9002\u5e94\u5404\u79cd\u786c\u4ef6\u548c\u90e8\u7f72\u8981\u6c42\u3002<\/p>\n<\/li>\n<li>\n<p><strong>\u53ef\u6269\u5c55\u6027\uff1a<\/strong> \u7528\u6237\u53ef\u4ee5\u5f00\u53d1\u81ea\u5b9a\u4e49\u5c42\u3001\u635f\u5931\u51fd\u6570\u548c\u5176\u4ed6\u7ec4\u4ef6\u6765\u6269\u5c55 Keras \u7684\u529f\u80fd\u3002<\/p>\n<\/li>\n<li>\n<p><strong>\u793e\u533a\u652f\u6301\uff1a<\/strong> Keras \u62e5\u6709\u4e00\u4e2a\u5145\u6ee1\u6d3b\u529b\u4e14\u6d3b\u8dc3\u7684\u793e\u533a\uff0c\u63d0\u4f9b\u5927\u91cf\u6587\u6863\u3001\u6559\u7a0b\u548c\u4ee3\u7801\u793a\u4f8b\u3002<\/p>\n<\/li>\n<\/ol>\n<h2>Keras \u7684\u7c7b\u578b<\/h2>\n<p>Keras \u6709\u4e0d\u540c\u7684\u5f62\u5f0f\u6765\u6ee1\u8db3\u4e0d\u540c\u7684\u9700\u6c42\u3002\u4ee5\u4e0b\u662f\u4e3b\u8981\u7c7b\u578b\uff1a<\/p>\n<ol>\n<li>\n<p><strong>\u72ec\u7acb Keras\uff1a<\/strong> Keras \u4e0e TensorFlow \u96c6\u6210\u4e4b\u524d\u7684\u539f\u59cb\u72ec\u7acb\u7248\u672c\u3002\u5b83\u4ecd\u7136\u53ef\u4ee5\u4f7f\u7528\uff0c\u4f46\u5927\u591a\u6570\u7528\u6237\u66f4\u559c\u6b22\u96c6\u6210\u7248\u672c\uff0c\u56e0\u4e3a TensorFlow \u5177\u6709\u66f4\u591a\u4f18\u52bf\u3002<\/p>\n<\/li>\n<li>\n<p><strong>TensorFlow \u4e2d\u7684 Keras API\uff1a<\/strong> \u8fd9\u662f Keras \u7684\u5b98\u65b9\u7248\u672c\uff0c\u96c6\u6210\u5230 TensorFlow \u5e93\u4e2d\u3002\u53ef\u4ee5\u901a\u8fc7\u4ee5\u4e0b\u65b9\u5f0f\u8bbf\u95ee\u5b83 <code data-no-translation=\"\">tf.keras<\/code> \u662f TensorFlow \u7528\u6237\u7684\u63a8\u8350\u9009\u62e9\u3002<\/p>\n<\/li>\n<\/ol>\n<h2>Keras\u7684\u4f7f\u7528\u65b9\u6cd5\u3001\u4f7f\u7528\u8fc7\u7a0b\u4e2d\u9047\u5230\u7684\u95ee\u9898\u4ee5\u53ca\u89e3\u51b3\u65b9\u6cd5<\/h2>\n<p>Keras \u53ef\u4ee5\u901a\u8fc7\u591a\u79cd\u65b9\u5f0f\u4f7f\u7528\uff0c\u5177\u4f53\u53d6\u51b3\u4e8e\u6df1\u5ea6\u5b66\u4e60\u9879\u76ee\u7684\u590d\u6742\u6027\u548c\u89c4\u6a21\u3002\u4e00\u4e9b\u5e38\u89c1\u7684\u4f7f\u7528\u573a\u666f\u5305\u62ec\uff1a<\/p>\n<ol>\n<li>\n<p><strong>\u521d\u5b66\u8005\u9879\u76ee\uff1a<\/strong> \u5bf9\u4e8e\u521d\u5b66\u8005\u6765\u8bf4\uff0cKeras \u63d0\u4f9b\u4e86\u4e00\u79cd\u5728\u5c0f\u578b\u6570\u636e\u96c6\u4e0a\u5b9e\u73b0\u548c\u8bad\u7ec3\u57fa\u672c\u795e\u7ecf\u7f51\u7edc\uff08\u4f8b\u5982\u524d\u9988\u6216\u5377\u79ef\u795e\u7ecf\u7f51\u7edc\uff09\u7684\u7b80\u5355\u65b9\u6cd5\u3002<\/p>\n<\/li>\n<li>\n<p><strong>\u7814\u7a76\u548c\u539f\u578b\u8bbe\u8ba1\uff1a<\/strong> \u7531\u4e8e Keras \u7684\u6613\u7528\u6027\u548c\u5feb\u901f\u6a21\u578b\u8fed\u4ee3\u529f\u80fd\uff0c\u7814\u7a76\u4eba\u5458\u548c\u6570\u636e\u79d1\u5b66\u5bb6\u7ecf\u5e38\u5229\u7528 Keras \u8fdb\u884c\u5feb\u901f\u539f\u578b\u8bbe\u8ba1\u548c\u5b9e\u9a8c\u3002<\/p>\n<\/li>\n<li>\n<p><strong>\u8fc1\u79fb\u5b66\u4e60\uff1a<\/strong> Keras \u4fc3\u8fdb\u4e86\u9884\u8bad\u7ec3\u6a21\u578b\u548c\u8fc1\u79fb\u5b66\u4e60\u7684\u4f7f\u7528\uff0c\u5176\u4e2d\u5728\u5927\u578b\u6570\u636e\u96c6\u4e0a\u8bad\u7ec3\u7684\u6a21\u578b\u9002\u7528\u4e8e\u7279\u5b9a\u4efb\u52a1\u3002<\/p>\n<\/li>\n<li>\n<p><strong>\u751f\u4ea7\u7ea7\u5e94\u7528\uff1a<\/strong> \u5bf9\u4e8e\u5927\u89c4\u6a21\u751f\u4ea7\u90e8\u7f72\uff0cKeras \u4e0e TensorFlow \u96c6\u6210\uff0c\u53ef\u901a\u8fc7 TensorFlow Serving \u6216 TensorFlow Lite \u5b9e\u73b0\u9ad8\u6548\u7684\u5206\u5e03\u5f0f\u8bad\u7ec3\u548c\u670d\u52a1\u3002<\/p>\n<\/li>\n<\/ol>\n<p>\u4e0e Keras \u4f7f\u7528\u76f8\u5173\u7684\u95ee\u9898\u901a\u5e38\u6d89\u53ca\u4e0d\u540c\u540e\u7aef\u7684\u517c\u5bb9\u6027\u95ee\u9898\u6216\u9488\u5bf9\u7279\u5b9a\u786c\u4ef6\u4f18\u5316\u6a21\u578b\u7684\u6311\u6218\u3002\u7136\u800c\uff0cKeras \u7684\u6d3b\u8dc3\u793e\u533a\u548c\u5e7f\u6cdb\u7684\u6587\u6863\u4e3a\u7528\u6237\u53ef\u80fd\u9047\u5230\u7684\u5927\u591a\u6570\u95ee\u9898\u63d0\u4f9b\u4e86\u89e3\u51b3\u65b9\u6848\u3002<\/p>\n<h2>\u4e3b\u8981\u7279\u70b9\u53ca\u4e0e\u540c\u7c7b\u672f\u8bed\u7684\u5176\u4ed6\u6bd4\u8f83<\/h2>\n<p>\u4e3a\u4e86\u66f4\u597d\u5730\u7406\u89e3 Keras \u7684\u610f\u4e49\uff0c\u6211\u4eec\u5c06\u5176\u4e0e\u7c7b\u4f3c\u7684\u6df1\u5ea6\u5b66\u4e60\u6846\u67b6\u8fdb\u884c\u6bd4\u8f83\uff1a<\/p>\n<table>\n<thead>\n<tr>\n<th>\u6846\u67b6<\/th>\n<th>\u4e3b\u8981\u7279\u5f81<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>\u5580\u62c9\u65af<\/td>\n<td>\u7528\u6237\u53cb\u597d\u3001\u6a21\u5757\u5316\u8bbe\u8ba1\u3001TensorFlow \u96c6\u6210\u3001\u7075\u6d3b\u6027\u548c\u5f3a\u5927\u7684\u793e\u533a\u652f\u6301\u3002<\/td>\n<\/tr>\n<tr>\n<td>TensorFlow<\/td>\n<td>\u591a\u529f\u80fd\u3001\u53ef\u6269\u5c55\u3001\u751f\u4ea7\u5c31\u7eea\u3001\u66f4\u5e7f\u6cdb\u7684\u751f\u6001\u7cfb\u7edf\uff0c\u5177\u6709\u5f3a\u5316\u5b66\u4e60\u5de5\u5177\u7b49\u3002<\/td>\n<\/tr>\n<tr>\n<td>\u706b\u70ac<\/td>\n<td>\u52a8\u6001\u8ba1\u7b97\u56fe\u3001\u5728\u7814\u7a76\u4e2d\u7684\u5e7f\u6cdb\u91c7\u7528\u3001\u66f4\u5bb9\u6613\u7684\u8c03\u8bd5\u4ee5\u53ca\u66f4\u591a\u7684 Python \u8bed\u6cd5\u3002<\/td>\n<\/tr>\n<tr>\n<td>\u5496\u5561<\/td>\n<td>\u4e13\u95e8\u9488\u5bf9\u8ba1\u7b97\u673a\u89c6\u89c9\u4efb\u52a1\uff0c\u63a8\u7406\u901f\u5ea6\u66f4\u5feb\uff0c\u4f46\u6a21\u578b\u5b9a\u5236\u7075\u6d3b\u6027\u8f83\u5dee\u3002<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<p>\u4e0e\u5176\u4ed6\u6846\u67b6\u76f8\u6bd4\uff0cKeras \u4ee5\u5176\u7528\u6237\u53cb\u597d\u6027\u548c\u6613\u7528\u6027\u800c\u8457\u79f0\uff0c\u4f7f\u5176\u6210\u4e3a\u521d\u5b66\u8005\u548c\u4e13\u6ce8\u4e8e\u5feb\u901f\u539f\u578b\u8bbe\u8ba1\u7684\u4eba\u7684\u9996\u9009\u3002<\/p>\n<h2>\u4e0e Keras \u76f8\u5173\u7684\u672a\u6765\u524d\u666f\u548c\u6280\u672f<\/h2>\n<p>Keras \u7684\u672a\u6765\u4e0e\u6df1\u5ea6\u5b66\u4e60\u53ca\u5176\u5e94\u7528\u7684\u53d1\u5c55\u5bc6\u5207\u76f8\u5173\u3002\u968f\u7740\u6df1\u5ea6\u5b66\u4e60\u7684\u4e0d\u65ad\u53d1\u5c55\uff0cKeras \u9884\u8ba1\u5c06\u91c7\u7528\u65b0\u6280\u672f\u548c\u67b6\u6784\u6765\u4fdd\u6301\u76f8\u5173\u6027\u3002 Keras \u7684\u4e00\u4e9b\u6f5c\u5728\u672a\u6765\u53d1\u5c55\u5305\u62ec\uff1a<\/p>\n<ol>\n<li>\n<p><strong>\u6539\u8fdb\u7684\u6027\u80fd\uff1a<\/strong> Keras \u53ef\u80fd\u4f1a\u53d7\u76ca\u4e8e\u6301\u7eed\u7684\u4f18\u5316\u5de5\u4f5c\uff0c\u4ece\u800c\u80fd\u591f\u5728\u5404\u79cd\u786c\u4ef6\u67b6\u6784\u4e0a\u5b9e\u73b0\u66f4\u5feb\u7684\u8bad\u7ec3\u548c\u63a8\u7406\u3002<\/p>\n<\/li>\n<li>\n<p><strong>\u81ea\u52a8\u673a\u5668\u5b66\u4e60\u96c6\u6210\uff1a<\/strong> Keras \u4e2d\u81ea\u52a8\u5316\u673a\u5668\u5b66\u4e60 (AutoML) \u6280\u672f\u7684\u96c6\u6210\u53ef\u4ee5\u4f7f\u7528\u6237\u80fd\u591f\u81ea\u52a8\u641c\u7d22\u6700\u4f73\u795e\u7ecf\u7f51\u7edc\u67b6\u6784\u3002<\/p>\n<\/li>\n<li>\n<p><strong>\u5bf9\u65b0\u67b6\u6784\u7684\u652f\u6301\uff1a<\/strong> \u968f\u7740\u65b0\u578b\u795e\u7ecf\u7f51\u7edc\u67b6\u6784\u7684\u51fa\u73b0\uff0cKeras \u9884\u8ba1\u5c06\u652f\u6301\u8fd9\u4e9b\u6a21\u578b\uff0c\u8fdb\u4e00\u6b65\u6269\u5c55\u5176\u5728\u5404\u4e2a\u9886\u57df\u7684\u9002\u7528\u6027\u3002<\/p>\n<\/li>\n<li>\n<p><strong>\u6301\u7eed\u7814\u7a76\u5408\u4f5c\uff1a<\/strong> Keras \u53ef\u80fd\u4f1a\u4fdd\u6301\u4e0e TensorFlow \u793e\u533a\u7684\u5bc6\u5207\u5408\u4f5c\uff0c\u4ece\u8be5\u9886\u57df\u7684\u8fdb\u6b65\u4e2d\u83b7\u76ca\u5e76\u505a\u51fa\u8d21\u732e\u3002<\/p>\n<\/li>\n<\/ol>\n<h2>\u5982\u4f55\u4f7f\u7528\u4ee3\u7406\u670d\u52a1\u5668\u6216\u5c06\u4ee3\u7406\u670d\u52a1\u5668\u4e0e Keras \u5173\u8054<\/h2>\n<p>\u4ee3\u7406\u670d\u52a1\u5668\u53ef\u4ee5\u5728\u589e\u5f3a Keras \u5e94\u7528\u7a0b\u5e8f\u7684\u6027\u80fd\u65b9\u9762\u53d1\u6325\u4f5c\u7528\uff0c\u7279\u522b\u662f\u5728\u6570\u636e\u8bbf\u95ee\u6216\u6a21\u578b\u670d\u52a1\u7531\u4e8e\u5730\u7406\u6216\u7f51\u7edc\u9650\u5236\u800c\u53d7\u5230\u9650\u5236\u7684\u60c5\u51b5\u4e0b\u3002\u4ee5\u4e0b\u662f\u4ee3\u7406\u670d\u52a1\u5668\u4e0e Keras \u4e00\u8d77\u4f7f\u7528\u7684\u4e00\u4e9b\u65b9\u6cd5\uff1a<\/p>\n<ol>\n<li>\n<p><strong>\u6570\u636e\u8bbf\u95ee\uff1a<\/strong> \u5728\u67d0\u4e9b\u60c5\u51b5\u4e0b\uff0c\u7528\u4e8e\u8bad\u7ec3\u6df1\u5ea6\u5b66\u4e60\u6a21\u578b\u7684\u6570\u636e\u53ef\u80fd\u5206\u5e03\u5728\u4e0d\u540c\u7684\u5730\u7406\u4f4d\u7f6e\u3002\u4ee3\u7406\u670d\u52a1\u5668\u53ef\u4ee5\u901a\u8fc7\u7f13\u5b58\u548c\u52a0\u901f\u6570\u636e\u8bbf\u95ee\u6765\u5b9e\u73b0\u9ad8\u6548\u7684\u6570\u636e\u68c0\u7d22\uff0c\u4ece\u800c\u7f29\u77ed\u8bad\u7ec3\u65f6\u95f4\u3002<\/p>\n<\/li>\n<li>\n<p><strong>\u8d1f\u8f7d\u5747\u8861\uff1a<\/strong> \u5728\u9ad8\u6d41\u91cf\u573a\u666f\u4e2d\uff0c\u90e8\u7f72\u4ee3\u7406\u670d\u52a1\u5668\u53ef\u4ee5\u5e2e\u52a9\u5c06\u4f20\u5165\u8bf7\u6c42\u5206\u53d1\u5230\u591a\u4e2a Keras \u652f\u6301\u7684\u670d\u52a1\u5668\uff0c\u786e\u4fdd\u8ba1\u7b97\u8d44\u6e90\u7684\u9ad8\u6548\u5229\u7528\u5e76\u7f29\u77ed\u54cd\u5e94\u65f6\u95f4\u3002<\/p>\n<\/li>\n<li>\n<p><strong>\u5b89\u5168\u548c\u9690\u79c1\uff1a<\/strong> \u4ee3\u7406\u670d\u52a1\u5668\u53ef\u4ee5\u5145\u5f53\u7528\u6237\u548c Keras \u5e94\u7528\u7a0b\u5e8f\u4e4b\u95f4\u7684\u4e2d\u4ecb\uff0c\u901a\u8fc7\u5c4f\u853d\u5b9e\u9645\u670d\u52a1\u5668 IP \u5e76\u4fdd\u62a4\u654f\u611f\u6570\u636e\u6765\u589e\u52a0\u989d\u5916\u7684\u5b89\u5168\u5c42\u3002<\/p>\n<\/li>\n<li>\n<p><strong>\u5185\u5bb9\u8fc7\u6ee4\uff1a<\/strong> \u4ee3\u7406\u670d\u52a1\u5668\u53ef\u4ee5\u8fc7\u6ee4\u548c\u9650\u5236\u5bf9\u67d0\u4e9b\u5185\u5bb9\u7684\u8bbf\u95ee\uff0c\u8fd9\u5bf9\u4e8e\u63a7\u5236\u5bf9 Keras \u6a21\u578b\u7684\u8bbf\u95ee\u6216\u6839\u636e\u7528\u6237\u8981\u6c42\u63d0\u4f9b\u7279\u5b9a\u8f93\u51fa\u975e\u5e38\u6709\u7528\u3002<\/p>\n<\/li>\n<\/ol>\n<h2>\u76f8\u5173\u94fe\u63a5<\/h2>\n<p>\u6709\u5173 Keras \u7684\u66f4\u591a\u4fe1\u606f\uff0c\u60a8\u53ef\u4ee5\u63a2\u7d22\u4ee5\u4e0b\u8d44\u6e90\uff1a<\/p>\n<ul>\n<li><a href=\"https:\/\/keras.io\/\" target=\"_new\" rel=\"noopener nofollow\">Keras \u6587\u6863<\/a><\/li>\n<li><a href=\"https:\/\/github.com\/keras-team\/keras\" target=\"_new\" rel=\"noopener nofollow\">Keras GitHub \u5b58\u50a8\u5e93<\/a><\/li>\n<li><a href=\"https:\/\/www.tensorflow.org\/\" target=\"_new\" rel=\"noopener nofollow\">TensorFlow \u5b98\u65b9\u7f51\u7ad9<\/a><\/li>\n<li><a href=\"https:\/\/pytorch.org\/\" target=\"_new\" rel=\"noopener nofollow\">PyTorch \u5b98\u65b9\u7f51\u7ad9<\/a><\/li>\n<li><a href=\"http:\/\/caffe.berkeleyvision.org\/\" target=\"_new\" rel=\"noopener nofollow\">\u5496\u5561\u5b98\u65b9\u7f51\u7ad9<\/a><\/li>\n<\/ul>\n<p>\u603b\u4e4b\uff0cKeras \u5df2\u6210\u4e3a\u9886\u5148\u7684\u6df1\u5ea6\u5b66\u4e60\u6846\u67b6\uff0c\u56e0\u5176\u7528\u6237\u53cb\u597d\u6027\u548c\u5f3a\u5927\u7684\u529f\u80fd\u800c\u53d7\u5230\u8d5e\u8d4f\u3002\u5b83\u4e0e TensorFlow \u7684\u65e0\u7f1d\u96c6\u6210\u4e3a\u7528\u6237\u63d0\u4f9b\u4e86\u4e00\u4e2a\u5f3a\u5927\u800c\u7075\u6d3b\u7684\u5e73\u53f0\u6765\u6784\u5efa\u548c\u90e8\u7f72\u795e\u7ecf\u7f51\u7edc\u3002\u968f\u7740\u6df1\u5ea6\u5b66\u4e60\u9886\u57df\u7684\u4e0d\u65ad\u53d1\u5c55\uff0cKeras \u6709\u671b\u968f\u4e4b\u53d1\u5c55\uff0c\u4fdd\u6301\u5728\u4eba\u5de5\u667a\u80fd\u548c\u673a\u5668\u5b66\u4e60\u521b\u65b0\u7684\u524d\u6cbf\u3002<\/p>","protected":false},"featured_media":468725,"menu_order":0,"template":"","meta":{"_acf_changed":false,"content-type":"","inline_featured_image":false,"footnotes":""},"class_list":["post-477761","wiki","type-wiki","status-publish","has-post-thumbnail","hentry"],"acf":{"faq_title":"Frequently Asked Questions about <mark>Keras: A Comprehensive Guide<\/mark>","faq_items":[{"question":"What is Keras?","answer":"<p>Keras is an open-source deep learning framework written in Python. It is known for its user-friendliness and flexibility, making it a popular choice among researchers, developers, and data scientists for building and experimenting with neural networks.<\/p>"},{"question":"Who developed Keras and when was it released?","answer":"<p>Keras was developed by Fran\u00e7ois Chollet and was first released in March 2015.<\/p>"},{"question":"What are the key features of Keras?","answer":"<p>Keras offers several key features, including a user-friendly API, modularity for building complex neural networks, seamless integration with TensorFlow, and extensibility to customize models.<\/p>"},{"question":"What types of Keras are there?","answer":"<p>There are two main types of Keras: the standalone version, which existed before integration with TensorFlow, and the integrated version, known as <code>tf.keras<\/code>, which is the official version integrated into the TensorFlow library.<\/p>"},{"question":"How does Keras work internally?","answer":"<p>Keras follows a high-level API design, with three main components: the backend for executing computational operations, layers for building neural network components, and models to organize the layers into a cohesive structure.<\/p>"},{"question":"How can proxy servers be associated with Keras?","answer":"<p>Proxy servers can enhance the performance of Keras applications by facilitating efficient data retrieval, load balancing, security, and privacy measures, and content filtering.<\/p>"},{"question":"What are the future perspectives of Keras?","answer":"<p>The future of Keras is expected to see improved performance, potential integration with AutoML techniques, support for new architectures, and continued collaboration with the TensorFlow community.<\/p>"},{"question":"Where can I find more information about Keras?","answer":"<p>For more information about Keras, you can explore the official <a href=\"https:\/\/keras.io\/\" target=\"_new\">Keras documentation<\/a>, visit the <a href=\"https:\/\/github.com\/keras-team\/keras\" target=\"_new\">Keras GitHub repository<\/a>, and also check out the official websites of <a href=\"https:\/\/www.tensorflow.org\/\" target=\"_new\">TensorFlow<\/a>, <a href=\"https:\/\/pytorch.org\/\" target=\"_new\">PyTorch<\/a>, and <a href=\"http:\/\/caffe.berkeleyvision.org\/\" target=\"_new\">Caffe<\/a>.<\/p>"}]},"_links":{"self":[{"href":"https:\/\/oneproxy.pro\/cn\/wp-json\/wp\/v2\/wiki\/477761","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/oneproxy.pro\/cn\/wp-json\/wp\/v2\/wiki"}],"about":[{"href":"https:\/\/oneproxy.pro\/cn\/wp-json\/wp\/v2\/types\/wiki"}],"version-history":[{"count":0,"href":"https:\/\/oneproxy.pro\/cn\/wp-json\/wp\/v2\/wiki\/477761\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/oneproxy.pro\/cn\/wp-json\/wp\/v2\/media\/468725"}],"wp:attachment":[{"href":"https:\/\/oneproxy.pro\/cn\/wp-json\/wp\/v2\/media?parent=477761"}],"curies":[{"name":"\u53ef\u6e7f\u6027\u7c89\u5242","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}