{"id":9908,"date":"2023-01-29T11:51:48","date_gmt":"2023-01-29T10:51:48","guid":{"rendered":"https:\/\/geehootek.com\/?p=9908"},"modified":"2023-01-29T11:51:50","modified_gmt":"2023-01-29T10:51:50","slug":"ai-and-machine-learning-for-coders-%e4%b8%ad%e6%96%87%e7%bf%bb%e8%af%91%ef%bc%882%ef%bc%89","status":"publish","type":"post","link":"https:\/\/geehootek.com\/zh-hans\/ai-and-machine-learning-for-coders-%e4%b8%ad%e6%96%87%e7%bf%bb%e8%af%91%ef%bc%882%ef%bc%89\/","title":{"rendered":"AI and machine learning for coders \u4e2d\u6587\u7ffb\u8bd1\uff082\uff09"},"content":{"rendered":"\n

\u4e2d\u6587\u4e66\u540d\uff1a\u7ed9\u7a0b\u5e8f\u5458\u7684\u4eba\u5de5\u667a\u80fd\u548c\u673a\u5668\u5b66\u4e60<\/p>\n\n\n\n

\u8bd1\u8005\uff1a\u6731\u745e<\/em><\/p>\n\n\n\n

\u8bd1\u8005\u90ae\u7bb1\uff1akevindree@gmail.com<\/em><\/p>\n\n\n\n

\u8bd1\u8005\u58f0\u660e\uff1a\u8bd1\u8005\u5728IT\u9886\u57df\u5de5\u4f5c\u591a\u5e74\uff0c\u57fa\u4e8e\u5bf9\u4e8e\u4eba\u5de5\u667a\u80fd\u3001\u673a\u5668\u5b66\u4e60\u7684\u5174\u8da3\uff0c\u81ea\u5b66\u4e86\u5f88\u591a\u673a\u5668\u5b66\u4e60\u8bfe\u7a0b\uff0cAi and machine learning for coders\u662f\u5176\u4e2d\u7684\u4e00\u672c\u4e66\u3002\u4e3a\u4e86\u8ba9\u66f4\u591a\u534e\u4eba\u4e86\u89e3\u548c\u5b66\u4e60\u4eba\u5de5\u667a\u80fd\u7684\u76f8\u5173\u6280\u672f\uff0c\u5728\u5b66\u4e60\u7684\u8fc7\u7a0b\u4e2d\uff0c\u7ecf\u8fc7\u957f\u65f6\u95f4\u7684\u8f9b\u52e4\u52b3\u52a8\u548c\u4ed8\u51fa\uff0c\u5c06\u8fd9\u4e00\u4e2d\u6587\u7248\u5448\u73b0\u5728\u5927\u5bb6\u9762\u524d\u3002\u7ffb\u8bd1\u4e2d\u96be\u514d\u6709\u8bb8\u591a\u4e0d\u51c6\u786e\u7684\u5730\u65b9\uff0c\u8fd8\u671b\u5e7f\u5927\u8bfb\u8005\u539f\u8c05\u3002<\/em><\/p>\n\n\n\n

Geehootek.com<\/a>\u662f\u8bd1\u8005\u9996\u53d1\u7f51\u7ad9\uff0c\u8bf7\u5c0a\u91cd\u8bd1\u8005\u52b3\u52a8\uff0c\u8f6c\u8f7d\u65f6\u8bf7\u6ce8\u660e\u51fa\u5904\uff01\uff01\uff01<\/mark><\/em><\/mark><\/p>\n\n\n\n

\u4ec0\u4e48\u662fTensorFlow<\/h4>\n\n\n\n

TensorFlow \u662f\u4e00\u4e2a\u7528\u4e8e\u521b\u5efa\u548c\u4f7f\u7528\u673a\u5668\u5b66\u4e60\u6a21\u578b\u7684\u5f00\u6e90\u5e73\u53f0\u3002 \u5b83\u5b9e\u73b0\u4e86\u673a\u5668\u5b66\u4e60\u6240\u9700\u7684\u8bb8\u591a\u5e38\u7528\u7b97\u6cd5\u548c\u6a21\u5f0f\uff0c\u4f7f\u60a8\u65e0\u9700\u5b66\u4e60\u6240\u6709\u5e95\u5c42\u6570\u5b66\u548c\u903b\u8f91\uff0c\u5e76\u4f7f\u60a8\u80fd\u591f\u4e13\u6ce8\u4e8e\u60a8\u7684\u573a\u666f\u3002 \u5b83\u9762\u5411\u6240\u6709\u4eba\uff0c\u4ece\u7231\u597d\u8005\u5230\u4e13\u4e1a\u5f00\u53d1\u4eba\u5458\uff0c\u518d\u5230\u63a8\u52a8\u4eba\u5de5\u667a\u80fd\u53d1\u5c55\u7684\u7814\u7a76\u4eba\u5458\u3002 \u91cd\u8981\u7684\u662f\uff0c\u5b83\u8fd8\u652f\u6301\u5c06\u6a21\u578b\u90e8\u7f72\u5230 Web\u3001\u4e91\u3001\u79fb\u52a8\u548c\u5d4c\u5165\u5f0f\u7cfb\u7edf\u3002 \u6211\u4eec\u5c06\u5728\u672c\u4e66\u4e2d\u6db5\u76d6\u6240\u6709\u8fd9\u4e9b\u573a\u666f\u3002<\/p>\n\n\n\n

TensorFlow \u7684\u9ad8\u5c42\u67b6\u6784\u5982\u56fe 1-11 \u6240\u793a\u3002<\/p>\n\n\n\n

\"\"
Figure 1-11. TensorFlow high-level architecture<\/figcaption><\/figure>\n\n\n\n

\u521b\u5efa\u673a\u5668\u5b66\u4e60\u6a21\u578b\u7684\u8fc7\u7a0b\u79f0\u4e3a\u8bad\u7ec3<\/em>\u3002 \u8fd9\u6307\u7684\u662f\u8ba1\u7b97\u673a\u4f7f\u7528\u4e00\u7ec4\u7b97\u6cd5\u6765\u5b66\u4e60\u8f93\u5165\u4ee5\u53ca\u4e0d\u540c\u8f93\u5165\u4e4b\u95f4\u7684\u533a\u522b\u7684\u65b9\u6cd5\u3002\u4f8b\u5982\uff0c\u5982\u679c\u4f60\u60f3\u8ba9\u8ba1\u7b97\u673a\u6765\u8bc6\u522b\u732b\u548c\u72d7\uff0c\u4f60\u53ef\u4ee5\u4f7f\u7528\u4e24\u8005\u7684\u5927\u91cf\u56fe\u7247\u6765\u521b\u5efa\u6a21\u578b\uff0c\u8ba1\u7b97\u673a\u5c06\u4f7f\u7528\u8be5\u6a21\u578b\u6765\u5c1d\u8bd5\u5f04\u6e05\u695a\u662f\u4ec0\u4e48\u662f\u732b\uff0c\u4ec0\u4e48\u662f\u72d7\u3002\u6a21\u578b\u5b8c\u6210\u8bad\u7ec3\u540e\uff0c\u8ba9\u5b83\u5bf9\u672a\u6765\u7684\u8f93\u5165\u8fdb\u884c\u8bc6\u522b\u6216\u5206\u7c7b\u7684\u8fc7\u7a0b\u79f0\u4e3a\u63a8\u7406<\/em>\u3002<\/p>\n\n\n\n

\u56e0\u6b64\uff0c\u8981\u8bad\u7ec3\u6a21\u578b\uff0c\u60a8\u9700\u8981\u51e0\u9879\u5185\u5bb9\u3002 \u9996\u5148\u662f\u4e00\u7ec4\u7528\u4e8e\u8bbe\u8ba1\u6a21\u578b\u7684 API\u3002 \u4f7f\u7528 TensorFlow \u53ef\u4ee5\u901a\u8fc7\u4e09\u79cd\u4e3b\u8981\u65b9\u5f0f\u6765\u505a\u5230\u8fd9\u4e00\u70b9\uff1a\u60a8\u53ef\u4ee5\u624b\u52a8\u7f16\u5199\u6240\u6709\u4ee3\u7801\uff0c\u627e\u51fa\u8ba1\u7b97\u673a\u5982\u4f55\u5b66\u4e60\u7684\u903b\u8f91\uff0c\u7136\u540e\u5728\u4ee3\u7801\u4e2d\u5b9e\u73b0\uff08\u4e0d\u63a8\u8350\uff09\uff1b \u60a8\u53ef\u4ee5\u4f7f\u7528\u5185\u7f6e\u7684\u4f30\u8ba1\u5668\uff0c\u5b83\u4eec\u662f\u5df2\u7ecf\u5b9e\u73b0\u7684\u795e\u7ecf\u7f51\u7edc\uff0c\u60a8\u53ef\u4ee5\u5bf9\u5176\u8fdb\u884c\u81ea\u5b9a\u4e49\uff1b \u6216\u8005\u60a8\u53ef\u4ee5\u4f7f\u7528 Keras\uff0c\u8fd9\u662f\u4e00\u79cd\u9ad8\u7ea7 API\uff0c\u53ef\u8ba9\u60a8\u5728\u4ee3\u7801\u4e2d\u5c01\u88c5\u5e38\u89c1\u7684\u673a\u5668\u5b66\u4e60\u8303\u4f8b\u3002 \u672c\u4e66\u5c06\u4e3b\u8981\u5173\u6ce8\u5728\u521b\u5efa\u6a21\u578b\u65f6\u4f7f\u7528 Keras API\u3002<\/p>\n\n\n\n

\u8bad\u7ec3\u6a21\u578b\u7684\u65b9\u6cd5\u6709\u5f88\u591a\u79cd\u3002 \u5728\u5927\u591a\u6570\u60c5\u51b5\u4e0b\uff0c\u60a8\u53ef\u80fd\u53ea\u4f7f\u7528\u5355\u4e2a\u82af\u7247\uff0c\u65e0\u8bba\u662f\u4e2d\u592e\u5904\u7406\u5668 (CPU)\u3001\u56fe\u5f62\u5904\u7406\u5668 (GPU) \u8fd8\u662f\u79f0\u4e3a\u5f20\u91cf\u5904\u7406\u5668 (TPU) \u7684\u65b0\u4e1c\u897f\u3002 \u5728\u66f4\u9ad8\u7ea7\u7684\u5de5\u4f5c\u548c\u7814\u7a76\u73af\u5883\u4e2d\uff0c\u53ef\u4ee5\u4f7f\u7528\u8de8\u591a\u4e2a\u82af\u7247\u7684\u5e76\u884c\u8bad\u7ec3\uff0c\u91c7\u7528\u8de8\u591a\u4e2a\u82af\u7247\u8bad\u7ec3\u7684\u5206\u5e03\u7b56\u7565\u3002 TensorFlow \u4e5f\u652f\u6301\u8fd9\u4e00\u70b9\u3002<\/p>\n\n\n\n

\u4efb\u4f55\u6a21\u578b\u7684\u751f\u547d\u7ebf\u90fd\u662f\u5b83\u7684\u6570\u636e\u3002 \u5982\u524d\u6240\u8ff0\uff0c\u5982\u679c\u60a8\u60f3\u521b\u5efa\u4e00\u4e2a\u53ef\u4ee5\u8bc6\u522b\u732b\u72d7\u7684\u6a21\u578b\uff0c\u5219\u9700\u8981\u4f7f\u7528\u5927\u91cf\u732b\u72d7\u7684\u793a\u4f8b\u5bf9\u5176\u8fdb\u884c\u8bad\u7ec3\u3002 \u4f46\u662f\u60a8\u5982\u4f55\u7ba1\u7406\u8fd9\u4e9b\u793a\u4f8b\u5462\uff1f \u968f\u7740\u65f6\u95f4\u7684\u63a8\u79fb\uff0c\u60a8\u4f1a\u53d1\u73b0\u8fd9\u901a\u5e38\u6d89\u53ca\u6bd4\u521b\u5efa\u6a21\u578b\u672c\u8eab\u66f4\u591a\u7684\u7f16\u7801\u3002 TensorFlow \u9644\u5e26 API \u4ee5\u5c1d\u8bd5\u7b80\u5316\u6b64\u8fc7\u7a0b\uff0c\u79f0\u4e3a TensorFlow \u6570\u636e\u670d\u52a1\u3002 \u4e3a\u4e86\u65b9\u4fbf\u5b66\u4e60\uff0c\u5b83\u4eec\u5305\u62ec\u8bb8\u591a\u7ecf\u8fc7\u9884\u5904\u7406\u7684\u6570\u636e\u96c6\uff0c\u60a8\u53ef\u4ee5\u4ec5\u901a\u8fc7\u4e00\u884c\u4ee3\u7801\u4f7f\u7528\u4ed6\u4eec\u3002 \u5b83\u4eec\u8fd8\u4e3a\u60a8\u63d0\u4f9b\u4e86\u7528\u4e8e\u5904\u7406\u539f\u59cb\u6570\u636e\u7684\u5de5\u5177\uff0c\u4f7f\u5176\u66f4\u6613\u4e8e\u4f7f\u7528\u3002<\/p>\n\n\n\n

\u9664\u4e86\u521b\u5efa\u6a21\u578b\u4e4b\u5916\uff0c\u60a8\u8fd8\u9700\u8981\u80fd\u591f\u5c06\u5b83\u4eec\u4ea4\u4ed8\u7ed9\u4eba\u4eec\u4f7f\u7528\u3002 \u4e3a\u6b64\uff0cTensorFlow \u5305\u542b\u7528\u4e8e\u670d\u52a1\u7684 API\uff0c\u60a8\u53ef\u4ee5\u5728\u5176\u4e2d\u901a\u8fc7 HTTP \u8fde\u63a5\u4e3a\u4e91\u6216 Web \u7528\u6237\u63d0\u4f9b\u6a21\u578b\u63a8\u7406\u3002 \u5bf9\u4e8e\u5728\u79fb\u52a8\u8bbe\u5907\u6216\u5d4c\u5165\u5f0f\u7cfb\u7edf\u4e0a\u8fd0\u884c\u7684\u6a21\u578b\uff0c\u5219\u6709 TensorFlow Lite\uff0c\u5b83\u63d0\u4f9b\u4e86\u5728 Android \u548c iOS \u4ee5\u53ca\u57fa\u4e8e Linux \u7684\u5d4c\u5165\u5f0f\u7cfb\u7edf\uff08\u4f8b\u5982 Raspberry Pi\uff09\u4e0a\u8fdb\u884c\u6a21\u578b\u63a8\u7406\u7684\u5de5\u5177\u3002 TensorFlow Lite \u7684\u4e00\u4e2a\u5206\u652f\uff0c\u79f0\u4e3a TensorFlow Lite Micro (TFLM)\uff0c\u4e5f\u901a\u8fc7\u79f0\u4e3a TinyML \u7684\u65b0\u5174\u6982\u5ff5\u63d0\u4f9b\u5bf9\u5fae\u63a7\u5236\u5668\u7684\u63a8\u7406\u3002 \u6700\u540e\uff0c\u5982\u679c\u60a8\u60f3\u4e3a\u6d4f\u89c8\u5668\u6216 Node.js \u7528\u6237\u63d0\u4f9b\u6a21\u578b\uff0cTensorFlow.js \u4e5f\u63d0\u4f9b\u4e86\u4ee5\u8fd9\u79cd\u65b9\u5f0f\u8bad\u7ec3\u548c\u6267\u884c\u6a21\u578b\u7684\u80fd\u529b\u3002<\/p>\n\n\n\n

\u63a5\u4e0b\u6765\uff0c\u6211\u5c06\u5411\u60a8\u5c55\u793a\u5982\u4f55\u5b89\u88c5 TensorFlow\uff0c\u4ee5\u4fbf\u60a8\u53ef\u4ee5\u5f00\u59cb\u7528\u5b83\u521b\u5efa\u5e76\u4f7f\u7528\u673a\u5668\u5b66\u4e60\u6a21\u578b\u3002<\/p>\n\n\n\n

\u4f7f\u7528TensorFlow<\/h4>\n\n\n\n

\u5728\u672c\u8282\u4e2d\uff0c\u6211\u4eec\u5c06\u4ecb\u7ecd\u5b89\u88c5\u548c\u4f7f\u7528 TensorFlow \u7684\u4e09\u79cd\u4e3b\u8981\u65b9\u5f0f\u3002 \u6211\u4eec\u5c06\u4ece\u5982\u4f55\u4f7f\u7528\u547d\u4ee4\u884c\u5c06\u5176\u5b89\u88c5\u5230\u60a8\u7684\u5f00\u53d1\u73af\u5883\u5f00\u59cb\u3002 \u7136\u540e\uff0c\u6211\u4eec\u5c06\u63a2\u7d22\u4f7f\u7528\u6d41\u884c\u7684 PyCharm IDE\uff08\u96c6\u6210\u5f00\u53d1\u73af\u5883\uff09\u6765\u5b89\u88c5 TensorFlow\u3002 \u6700\u540e\uff0c\u6211\u4eec\u5c06\u4e86\u89e3 Google Colab \u4ee5\u53ca\u5982\u4f55\u4f7f\u7528\u5b83\u5728\u6d4f\u89c8\u5668\u4e2d\u901a\u8fc7\u57fa\u4e8e\u4e91\u7684\u540e\u7aef\u8bbf\u95ee TensorFlow \u4ee3\u7801\u3002<\/p>\n\n\n\n

\u5728 Python \u4e2d\u5b89\u88c5 TensorFlow<\/h5>\n\n\n\n

TensorFlow \u652f\u6301\u4f7f\u7528\u591a\u79cd\u8bed\u8a00\u521b\u5efa\u6a21\u578b\uff0c\u5305\u62ec Python\u3001Swift\u3001Java \u7b49\u3002 \u5728\u672c\u4e66\u4e2d\uff0c\u6211\u4eec\u5c06\u4e13\u6ce8\u4e8e\u4f7f\u7528 Python\uff0c\u7531\u4e8e\u5b83\u5bf9\u6570\u5b66\u6a21\u578b\u7684\u5e7f\u6cdb\u652f\u6301\uff0c\u5b83\u5b9e\u9645\u4e0a\u5df2\u7ecf\u6210\u4e3a\u673a\u5668\u5b66\u4e60\u7684\u8bed\u8a00\u3002 \u5982\u679c\u60a8\u8fd8\u6ca1\u6709python\uff0c\u6211\u5f3a\u70c8\u5efa\u8bae\u60a8\u8bbf\u95ee Python <\/a>\u4ee5\u542f\u52a8\u5e76\u8fd0\u884c\u5b83\uff0c\u5e76\u8bbf\u95ee learnpython.org<\/a> \u4ee5\u5b66\u4e60 Python \u8bed\u8a00\u8bed\u6cd5\u3002<\/p>\n\n\n\n

\u4f7f\u7528 Python \u6709\u591a\u79cd\u5b89\u88c5\u6846\u67b6\u7684\u65b9\u6cd5\uff0c\u4f46 TensorFlow \u56e2\u961f\u9ed8\u8ba4\u652f\u6301\u7684\u65b9\u6cd5\u662f pip\u3002<\/p>\n\n\n\n

\u56e0\u6b64\uff0c\u5728\u60a8\u7684 Python \u73af\u5883\u4e2d\uff0c\u5b89\u88c5 TensorFlow \u5341\u5206\u7b80\u5355\uff0c\u5982\u4e0b\uff1a<\/p>\n\n\n\n

pip install tensorflow<\/code><\/pre>\n\n\n\n
\u9700\u8981\u6ce8\u610f\u7684\u662f\uff0c\u4ece 2.1 \u7248\u672c\u5f00\u59cb\uff0c\u8fd9\u5c06\u9ed8\u8ba4\u5b89\u88c5 GPU \u7248\u672c\u7684 TensorFlow\u3002 \u5728\u6b64\u4e4b\u524d\uff0c\u5b83\u4f7f\u7528\u7684\u662f CPU \u7248\u672c\u3002 \u56e0\u6b64\uff0c\u5728\u5b89\u88c5\u4e4b\u524d\uff0c\u8bf7\u786e\u4fdd\u60a8\u62e5\u6709\u53d7\u652f\u6301\u7684 GPU \u53ca\u5176\u6240\u6709\u5fc5\u9700\u7684\u9a71\u52a8\u7a0b\u5e8f\u3002 \u6709\u5173\u8be6\u7ec6\u4fe1\u606f\uff0c\u8bf7\u8bbf\u95ee TensorFlow<\/a>\u3002<\/p>\n\n\n\n
\u5982\u679c\u60a8\u6ca1\u6709\u6240\u9700\u8981\u7684 GPU \u6216\u9a71\u52a8\u7a0b\u5e8f\uff0c\u60a8\u4ecd\u7136\u53ef\u4ee5\u9009\u62e9\u5728\u4efb\u4f55 Linux\u3001PC \u6216 Mac \u4e0a\u5b89\u88c5 TensorFlow \u7684 CPU \u7248\u672c\uff1a<\/p>\n\n\n\n
pip install tensorflow-cpu<\/code><\/pre>\n\n\n\n
\u5b8c\u6210\u5b89\u88c5\u540e\uff0c\u60a8\u53ef\u4ee5\u542f\u52a8\u5e76\u8fd0\u884cpython\uff0c\u5728python\u73af\u5883\u4e2d\u60a8\u53ef\u4ee5\u4f7f\u7528\u4ee5\u4e0b\u4ee3\u7801\u6d4b\u8bd5\u60a8\u7684 TensorFlow \u7248\u672c\uff1a<\/p>\n\n\n\n
import tensorflow as tf\r\nprint(tf.__version__)\r<\/code><\/pre>\n\n\n\n
\u4f60\u5e94\u8be5\u53ef\u4ee5\u770b\u5230\u5982\u56fe 1-12 \u6240\u793a\u7684\u8f93\u51fa\u7ed3\u679c\u3002 \u8fd9\u4e2a\u547d\u4ee4\u5c06\u6253\u5370\u5f53\u524d\u8fd0\u884c\u7684 TensorFlow \u7248\u672c\u2014\u2014\u5728\u8fd9\u91cc\u60a8\u53ef\u4ee5\u770b\u5230\u5df2\u5b89\u88c5\u7684\u7248\u672c\u662f 2.0.0\u3002<\/p>\n\n\n\n
\"\"
Figure 1-12. Running TensorFlow in Python<\/figcaption><\/figure>\n\n\n\n
\u5728 PyCharm \u4e2d\u4f7f\u7528 TensorFlow<\/h5>\n\n\n\n
\u6211\u7279\u522b\u559c\u6b22\u4f7f\u7528 PyCharm \u7684\u514d\u8d39\u793e\u533a\u7248\u672c\u6765\u7528 TensorFlow \u6784\u5efa\u6a21\u578b\u3002 PyCharm \u5f88\u597d\u7528\u7684\u539f\u56e0\u6709\u5f88\u591a\uff0c\u4f46\u6211\u6700\u559c\u6b22\u5b83\u7684\u539f\u56e0\u4e4b\u4e00\u662f\u5b83\u53ef\u4ee5\u975e\u5e38\u5bb9\u6613\u7684\u7ba1\u7406\u865a\u62df\u73af\u5883\u3002 \u8fd9\u610f\u5473\u7740\u60a8\u53ef\u4ee5\u62e5\u6709\u4e00\u4e2a\u5305\u542b\u9488\u5bf9\u60a8\u7279\u5b9a\u9879\u76ee\u7684\u7279\u5b9a\u5de5\u5177\u7248\u672c\uff08\u4f8b\u5982TensorFlow\uff09\u7684 Python \u73af\u5883\u3002 \u56e0\u6b64\uff0c\u5982\u679c\u4f60\u60f3\u5728\u4e00\u4e2a\u9879\u76ee\u4e2d\u4f7f\u7528 TensorFlow 2.0 \u800c\u5728\u53e6\u4e00\u4e2a\u9879\u76ee\u4e2d\u4f7f\u7528 TensorFlow 2.1\uff0c\u4f60\u53ef\u4ee5\u901a\u8fc7\u865a\u62df\u73af\u5883\u5c06\u5b83\u4eec\u5206\u5f00\uff0c\u800c\u4e0d\u5fc5\u5728\u5207\u6362\u73af\u5883\u65f6\u5904\u7406\u5b89\u88c5\/\u5378\u8f7d\u4f9d\u8d56\u9879\u3002 \u6b64\u5916\uff0c\u501f\u52a9 PyCharm\uff0c\u60a8\u53ef\u4ee5\u5bf9 Python \u4ee3\u7801\u8fdb\u884c\u5206\u6b65\u8c03\u8bd5\u2014\u2014\u8fd9\u662f\u5fc5\u987b\u7684\uff0c\u5c24\u5176\u662f\u5728\u60a8\u521a\u521a\u8d77\u6b65\u7684\u60c5\u51b5\u4e0b\u3002<\/p>\n\n\n\n
\u4f8b\u5982\uff0c\u5728\u56fe 1-13 \u4e2d\uff0c\u6211\u6709\u4e00\u4e2a\u540d\u4e3a example1 \u7684\u65b0\u9879\u76ee\uff0c\u6211\u6307\u5b9a\u8981\u4f7f\u7528 Conda \u521b\u5efa\u4e00\u4e2a\u65b0\u73af\u5883\u3002 \u5f53\u6211\u521b\u5efa\u9879\u76ee\u65f6\uff0c\u6211\u5c06\u62e5\u6709\u4e00\u4e2a\u5e72\u51c0\u7684\u65b0\u7684 Python \u865a\u62df\u73af\u5883\uff0c\u6211\u53ef\u4ee5\u5728\u5176\u4e2d\u5b89\u88c5\u6211\u60f3\u8981\u7684\u4efb\u4f55\u7248\u672c\u7684 TensorFlow\u3002<\/p>\n\n\n\n
\"\"
Figure 1-13. Creating a new virtual environment using PyCharm<\/figcaption><\/figure>\n\n\n\n
\u521b\u5efa\u9879\u76ee\u540e\uff0c\u60a8\u53ef\u4ee5\u6253\u5f00\u201cFile\u201d\u2192\u201cSetting\u201d\u5bf9\u8bdd\u6846\uff0c\u7136\u540e\u4ece\u5de6\u4fa7\u83dc\u5355\u4e2d\u9009\u62e9\u201cProject\uff1a<your proejct name>\u201d\u6761\u76ee\u3002 \u7136\u540e\u60a8\u5c06\u770b\u5230\u66f4\u6539\u9879\u76ee\u89e3\u91ca\u5668\u548c\u9879\u76ee\u7ed3\u6784\u8bbe\u7f6e\u7684\u9009\u9879\u3002 \u9009\u62e9 Project Interpreter \u94fe\u63a5\uff0c\u60a8\u5c06\u770b\u5230\u6b63\u5728\u4f7f\u7528\u7684\u89e3\u91ca\u5668\uff0c\u4ee5\u53ca\u5b89\u88c5\u5728\u8be5\u865a\u62df\u73af\u5883\u4e2d\u7684\u5305\u5217\u8868\uff08\u56fe 1-14\uff09\u3002<\/p>\n\n\n\n
\"\"
Figure 1-14. Adding packages to a virtual environment<\/figcaption><\/figure>\n\n\n\n
\u5355\u51fb\u53f3\u4fa7\u7684 + \u6309\u94ae\uff0c\u5c06\u6253\u5f00\u4e00\u4e2a\u5bf9\u8bdd\u6846\uff0c\u663e\u793a\u5f53\u524d\u53ef\u7528\u7684\u5305\u3002 \u5728\u641c\u7d22\u6846\u4e2d\u952e\u5165\u201ctensorflow\u201d\uff0c\u60a8\u5c06\u770b\u5230\u6240\u6709\u540d\u79f0\u4e2d\u5e26\u6709\u201ctensorflow\u201d\u7684\u53ef\u7528\u8f6f\u4ef6\u5305\uff08\u56fe 1-15\uff09\u3002<\/p>\n\n\n\n
\"\"
Figure 1-15. Installing TensorFlow with PyCharm<\/figcaption><\/figure>\n\n\n\n
\u5f53\u60a8\u9009\u62e9\u4e86 TensorFlow \u6216\u60a8\u60f3\u8981\u5b89\u88c5\u7684\u4efb\u4f55\u5176\u4ed6\u5305\uff0c\u5355\u51fb\u201cInstall Package\u201d\u6309\u94ae\uff0cPyCharm \u5c06\u5b8c\u6210\u5269\u4e0b\u7684\u5de5\u4f5c\u3002<\/p>\n\n\n\n
TensorFlow \u5b89\u88c5\u5b8c\u6210\u540e\uff0c\u60a8\u5c31\u53ef\u4ee5\u4f7f\u7528 Python \u7f16\u5199\u548c\u8c03\u8bd5 TensorFlow \u4ee3\u7801\u4e86\u3002<\/p>\n\n\n\n
\u5728 Google Colab \u4e2d\u4f7f\u7528 TensorFlow<\/h5>\n\n\n\n
\u53e6\u4e00\u4e2a\u53ef\u80fd\u6700\u5bb9\u6613\u4e0a\u624b\u7684\u9009\u62e9\u662f\u4f7f\u7528 Google Colab\uff0c\u8fd9\u662f\u4e00\u4e2a\u6258\u7ba1\u7684 Python \u73af\u5883\uff0c\u60a8\u53ef\u4ee5\u901a\u8fc7\u6d4f\u89c8\u5668\u8bbf\u95ee\u5b83\u3002 Colab \u7684\u771f\u6b63\u4f18\u70b9\u5728\u4e8e\u5b83\u63d0\u4f9b GPU \u548c TPU \u540e\u7aef\uff0c\u56e0\u6b64\u60a8\u53ef\u4ee5\u4f7f\u7528\u6700\u5148\u8fdb\u7684\u786c\u4ef6\u514d\u8d39\u8bad\u7ec3\u6a21\u578b\u3002<\/p>\n\n\n\n
\u5f53\u4f60\u8bbf\u95ee Colab \u7f51\u7ad9\u65f6\uff0c\u4f60\u53ef\u4ee5\u9009\u62e9\u6253\u5f00\u4ee5\u524d\u7684 Colabs \u6216\u5f00\u59cb\u4e00\u4e2a\u65b0\u7684\u7b14\u8bb0\u672c\uff0c\u5982\u56fe 1-16 \u6240\u793a\u3002<\/p>\n\n\n\n
\"\"
Figure 1-16. Getting started with Google Colab<\/figcaption><\/figure>\n\n\n\n
\u5355\u51fb New Python 3 Notebook \u94fe\u63a5\u5c06\u6253\u5f00\u7f16\u8f91\u5668\uff0c\u60a8\u53ef\u4ee5\u5728\u5176\u4e2d\u6dfb\u52a0\u4ee3\u7801\u6216\u6587\u672c\uff08\u56fe 1-17\uff09\u3002 \u60a8\u53ef\u4ee5\u901a\u8fc7\u5355\u51fb\u7a97\u683c\u5de6\u4fa7\u7684\u201c\u64ad\u653e\u201d\u6309\u94ae\uff08\u7bad\u5934\uff09\u6765\u6267\u884c\u4ee3\u7801\u3002<\/p>\n\n\n\n
\"\"
Figure 1-17. Running TensorFlow code in Colab<\/figcaption><\/figure>\n\n\n\n
\u5982\u6b64\u5904\u6240\u663e\u793a\u7684\uff0c\u68c0\u67e5 TensorFlow \u7248\u672c\u603b\u662f\u4e00\u4e2a\u597d\u7684\u4e60\u60ef\uff0c\u4ee5\u786e\u4fdd\u60a8\u8fd0\u884c\u7684\u662f\u6b63\u786e\u7684\u7248\u672c\u3002 Colab \u7684\u5185\u7f6e TensorFlow \u901a\u5e38\u4f1a\u6bd4\u6700\u65b0\u7248\u672c\u843d\u540e\u4e00\u4e24\u4e2a\u7248\u672c\u3002 \u5982\u679c\u662f\u8fd9\u79cd\u60c5\u51b5\uff0c\u60a8\u53ef\u4ee5\u4f7f\u7528 pip install \u66f4\u65b0\u5b83\uff0c\u53ea\u9700\u4f7f\u7528\u5982\u4e0b\u4ee3\u7801\u5757\uff1a<\/p>\n\n\n\n
!pip install tensorflow==2.1.0<\/p>\n\n\n\n
\u8fd0\u884c\u6b64\u547d\u4ee4\u540e\uff0c\u60a8\u5728 Colab \u4e2d\u7684\u5f53\u524d\u73af\u5883\u5c06\u4f7f\u7528\u6240\u9700\u7248\u672c\u7684 TensorFlow\u3002<\/p>\n","protected":false},"excerpt":{"rendered":"
\u4ec0\u4e48\u662fTensorFlow\uff1fTensorFlow \u662f\u4e00\u4e2a\u7528\u4e8e\u521b\u5efa\u548c\u4f7f\u7528\u673a\u5668\u5b66\u4e60\u6a21\u578b\u7684\u5f00\u6e90\u5e73\u53f0\u3002 \u5b83\u5b9e\u73b0\u4e86\u673a\u5668\u5b66\u4e60\u6240\u9700\u7684\u8bb8\u591a\u5e38\u7528\u7b97\u6cd5\u548c\u6a21\u5f0f\uff0c\u4f7f\u60a8\u65e0\u9700\u5b66\u4e60\u6240\u6709\u5e95\u5c42\u6570\u5b66\u548c\u903b\u8f91\uff0c\u5e76\u4f7f\u60a8\u80fd\u591f\u4e13\u6ce8\u4e8e\u60a8\u7684\u573a\u666f\u3002 \u5b83\u9762\u5411\u6240\u6709\u4eba\uff0c\u4ece\u7231\u597d\u8005\u5230\u4e13\u4e1a\u5f00\u53d1\u4eba\u5458\uff0c\u518d\u5230\u63a8\u52a8\u4eba\u5de5\u667a\u80fd\u53d1\u5c55\u7684\u7814\u7a76\u4eba\u5458\u3002 \u91cd\u8981\u7684\u662f\uff0c\u5b83\u8fd8\u652f\u6301\u5c06\u6a21\u578b\u90e8\u7f72\u5230 Web\u3001\u4e91\u3001\u79fb\u52a8<\/p>\n","protected":false},"author":1,"featured_media":9909,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[144],"tags":[],"class_list":["post-9908","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-144"],"aioseo_notices":[],"_links":{"self":[{"href":"https:\/\/geehootek.com\/zh-hans\/wp-json\/wp\/v2\/posts\/9908","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/geehootek.com\/zh-hans\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/geehootek.com\/zh-hans\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/geehootek.com\/zh-hans\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/geehootek.com\/zh-hans\/wp-json\/wp\/v2\/comments?post=9908"}],"version-history":[{"count":9,"href":"https:\/\/geehootek.com\/zh-hans\/wp-json\/wp\/v2\/posts\/9908\/revisions"}],"predecessor-version":[{"id":9942,"href":"https:\/\/geehootek.com\/zh-hans\/wp-json\/wp\/v2\/posts\/9908\/revisions\/9942"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/geehootek.com\/zh-hans\/wp-json\/wp\/v2\/media\/9909"}],"wp:attachment":[{"href":"https:\/\/geehootek.com\/zh-hans\/wp-json\/wp\/v2\/media?parent=9908"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/geehootek.com\/zh-hans\/wp-json\/wp\/v2\/categories?post=9908"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/geehootek.com\/zh-hans\/wp-json\/wp\/v2\/tags?post=9908"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}