![\"\"](\"https:\/\/www.zhuyanbin.com\/wp-content\/uploads\/2018\/08\/01-1.jpg\")
<\/a><\/p>\nTensorFlow\u4f7f\u7528\u6570\u636e\u6d41\u56fe\uff0c\u4e0a\u9762\u662f\u4e00\u4e2aTensorFlow\u7684\u793a\u610f\u56fe\u3002\u6211\u4eec\u4e0d\u7528\u5173\u5fc3\u8fd9\u5f20\u56fe\u5177\u4f53\u4ee3\u8868\u4ec0\u4e48\uff0c\u4f46\u4f60\u9700\u8981\u77e5\u9053\uff0c\u5176\u4e2d\u7684\u692d\u5706\u548c\u6b63\u65b9\u5f62\u4ee3\u8868\u8282\u70b9\uff0c\u8282\u70b9\u5c31\u662f\u4e00\u4e9b\u6570\u5b66\u51fd\u6570\uff0c\u4f60\u53ef\u4ee5\u5c06\u8282\u70b9\u5206\u7ec4\u5f62\u6210\u5404\u79cd\u6570\u5b66\u8ba1\u7b97\uff0c\u5e76\u5f97\u5230\u8f93\u51fa\u3002<\/p>\n
\u7bad\u5934\u4ee3\u8868\u6d41\u52a8\u5230\u5404\u4e2a\u8282\u70b9\u7684\u6570\u636e\u3002\u6240\u4ee5TensorFlow\u4e5f\u88ab\u6210\u4e3a\u6570\u636e\u6d41\u5e93\u3002<\/p>\n
OK\uff0c\u7b80\u4ecb\u5230\u6b64\uff0c\u4e0b\u9762\u5206\u4eab\u56db\u4e2a\u6211\u975e\u5e38\u559c\u6b22\u7684TensorFlow GitHub\u9879\u76ee\u3002<\/p>\n
\u9879\u76ee\u4e00\uff1aNeural Style<\/p>\n
\u8fd9\u662f\u6700\u9177\u7684TensorFlow GitHub\u9879\u76ee\u4e4b\u4e00\u3002\u795e\u7ecf\u98ce\u683c\u662f\u5c06\u4e00\u5f20\u7167\u7247\u7684\u98ce\u683c\u8fc1\u79fb\u5230\u53e6\u4e00\u5f20\u7167\u7247\u4e0a\u7684\u8fc7\u7a0b\uff0c\u540c\u65f6\u4fdd\u7559\u76f8\u5173\u7684\u7279\u5f20\u3002\u7b80\u5355\u7684\u6765\u8bf4\uff0c\u901a\u8fc7\u8fd9\u4e2a\u9879\u76ee\uff0c\u4f60\u53ef\u4ee5\u4f7f\u7528TensorFlow\u521b\u5efa\u81ea\u5df1\u7684Prisma\u5e94\u7528\u7a0b\u5e8f\u3002<\/p>\n
\u5982\u4e0a\u56fe\u6240\u793a\uff0c\u628a\u68b5\u9ad8\u753b\u4f5c\u7684\u98ce\u683c\uff0c\u8fc1\u79fb\u5230\u4e00\u5f20\u72ee\u5b50\u7684\u7167\u7247\u4e0a\uff0c\u5c31\u5f97\u5230\u4e00\u4e2a\u661f\u7a7a\u98ce\u683c\u7684\u72ee\u5b50\u7167\u7247\uff0c\u8fd9\u5c31\u662f\u6240\u8c13\u7684\u98ce\u683c\u8fc1\u79fb\u3002\u4e0b\u9762\u518d\u5c55\u793a\u4e00\u7ec4\u8fd9\u4e2a\u9879\u76ee\u7684\u98ce\u683c\u8fc1\u79fb\uff0c\u80fd\u628a\u8fd9\u5f20\u72ee\u5b50\u7684\u7167\u7247\uff0c\u53d8\u6210\u54ea\u4e9b\u975e\u5e38\u6709\u8da3\u7684\u6548\u679c\u3002<\/p>\n \u8fd9\u4e2a\u9879\u76ee\u5728\u98ce\u683c\u8fc1\u79fb\u7684\u540c\u65f6\uff0c\u8fd8\u80fd\u5bf9\u56fe\u50cf\u8fdb\u884c\u8bed\u4e49\u5206\u5272\u3002\u8fdb\u800c\u5b9e\u73b0\u5bf9\u5185\u5bb9\u7684\u524d\u666f\u3001\u80cc\u666f\u8fdb\u884c\u4e0d\u540c\u7684\u98ce\u683c\u8fc1\u79fb\u5904\u7406\u3002\u8fd9\u4e2a\u8fc7\u7a0b\u5982\u4e0b\u56fe\u6240\u793a\uff1a<\/p>\n \u53e6\u5916\u8fd9\u4e2a\u7b97\u6cd5\u8fd8\u80fd\u9002\u7528\u4e8e\u89c6\u9891\u5904\u7406\u3002<\/p>\n \u4e0a\u8ff0\u56fe\u7247\u5904\u7406\u4f7f\u7528\u7684\u8f6f\u786c\u4ef6\u5982\u4e0b\uff1a<\/p>\n CPU: Intel Core i7-6800K@3.4GHz\u00d712<\/p>\n GPU: NVIDIA GeForce GTX 1080\/PCIe\/SSE2<\/p>\n OS: Linux Ubuntu 16.04.1 LTS 64-bit<\/p>\n CUDA: 8.0<\/p>\n python: 2.7.12<\/p>\n tensorflow: 0.10.0rc<\/p>\n opencv: 2.4.9.1<\/p>\n GitHub\u5730\u5740\u5982\u4e0b\uff1a<\/p>\n https:\/\/github.com\/cysmith\/neural-style-tf<\/p>\n \u9879\u76ee\u4e8c\uff1aMozilla Deep Speech<\/p>\n \u8fd9\u4e2aGitHub\u9879\u76ee\u4f7f\u7528TensorFlow\u5c06\u8bed\u97f3\u8f6c\u6362\u4e3a\u6587\u672c\u3002\u8bed\u97f3\u8f6c\u6587\u672c\u662f\u4e00\u4e2a\u70ed\u95e8\u7684\u673a\u5668\u5b66\u4e60\u9886\u57df\uff0c\u7136\u800c\u5404\u5730\u7684\u4eba\u4eec\u6709\u7740\u4e0d\u540c\u7684\u53e3\u97f3\uff0c\u8fd9\u4e5f\u662f\u4e00\u4e2a\u96be\u4ee5\u89e3\u51b3\u7684\u95ee\u9898\u3002\u4e0d\u8fc7\u4ecd\u7136\u53ef\u4ee5\u901a\u8fc7\u6df1\u5ea6\u5b66\u4e60\u5b9e\u73b0\u975e\u5e38\u4e0d\u9519\u7684\u51c6\u786e\u6027\u3002<\/p>\n \u5176\u5b9e\u8fd9\u4e2a\u9879\u76ee\uff0c\u662f\u4e00\u4e2a\u57fa\u4e8e\u767e\u5ea6DeepSpeech\u67b6\u6784\u7684TensorFlow\u5b9e\u73b0\u3002\u6362\u53e5\u8bdd\u8bf4\uff0c\u8fd9\u4e2a\u9879\u76ee\u662f\u767e\u5ea6\u7684\u5185\u6838\uff0cGoogle\u7684\u5916\u8c8c\u3002<\/p>\n \u767e\u5ea6\u7684\u76f8\u5173\u8bba\u6587\u5730\u5740\u5728\u6b64\uff1a<\/p>\n https:\/\/arxiv.org\/abs\/1412.5567<\/p>\n DeepSpeech\u662f\u5434\u6069\u8fbe\u5e26\u9886\u767e\u5ea6\u56e2\u961f\u7814\u53d1\u51fa\u7684\u6210\u679c\uff0c\u6700\u65e9\u53d1\u5e03\u4e8e2014\u5e74\u5e95\u3002\u4eca\u5e74\u521d\uff0c\u767e\u5ea6\u57fa\u4e8eDeepSpeech2\uff0c\u5f00\u53d1\u51fa\u4e00\u6b3e\u540d\u4e3aSwiftScribe\u7684\u5e94\u7528\uff08swiftscribe.ai\uff09\uff0c\u53ef\u4ee5\u628a\u8bed\u97f3\u6587\u4ef6\u66f4\u4e3a\u5feb\u901f\u3001\u4fbf\u6377\u7684\u8f6c\u6362\u4e3a\u6587\u5b57\u3002\u76ee\u524d\u53ea\u9650\u4e8e\u82f1\u6587\u3002<\/p>\n \u56de\u5230\u8fd9\u4e2a\u9879\u76ee\u3002<\/p>\n \u6240\u9700\u8f6f\u4ef6\u73af\u5883\u5982\u4e0b\uff1a<\/p>\n Git Large File Storage<\/p>\n TensorFlow 1.0 or 1.1<\/p>\n SciPy<\/p>\n PyXDG<\/p>\n python_speech_features (nb: deprecated)<\/p>\n python sox<\/p>\n pandas<\/p>\n DeepSpeech native client libraries<\/p>\n \u5982\u679c\u4f60\u6709\u81f3\u5c118GB\u663e\u5b58\u7684\u82f1\u4f1f\u8fbeGPU\uff0c\u5f3a\u70c8\u5efa\u8bae\u5b89\u88c5\u652f\u6301GPU\u7684TensorFlow\uff0c\u56e0\u4e3a\u4f7f\u7528GPU\u7684\u8bad\u7ec3\u6bd4CPU\u5feb\u5f97\u591a\u3002<\/p>\n GitHub\u5730\u5740\u5982\u4e0b\uff1a<\/p>\n https:\/\/github.com\/mozilla\/DeepSpeech<\/p>\n \u9879\u76ee\u4e09\uff1a\u53e5\u5b50\u5206\u7c7b<\/p>\n \u53e5\u5b50\u5206\u7c7b\u5c31\u662f\u8bc6\u522b\u53e5\u5b50\u7c7b\u578b\u7684\u8fc7\u7a0b\u3002\u4f8b\u5982\uff0c\u5bf9\u4e8e\u201c\u98df\u7269\u975e\u5e38\u7cdf\u7cd5\u201d\u8fd9\u4e2a\u53e5\u5b50\uff0c\u4f60\u53ef\u80fd\u5e0c\u671b\u5c06\u5176\u5206\u7c7b\u4e3a\u6b63\u9762\u53e5\u5b50\u6216\u8d1f\u9762\u53e5\u5b50\uff0c\u8fd9\u4e5f\u88ab\u79f0\u4e3a\u60c5\u7eea\u5206\u6790\u3002\u8fd9\u4e2a\u95ee\u9898\u7684\u96be\u70b9\u5728\u4e8e\uff1a\u53e5\u5b50\u7ed3\u6784\u5e26\u6765\u7684\u590d\u6742\u6027\u3002<\/p>\n \u5229\u7528\u5377\u79ef\u795e\u7ecf\u7f51\u7edc\uff0c\u6211\u4eec\u53ef\u4ee5\u5c1d\u8bd5\u6784\u5efa\u4e00\u4e2a\u5f3a\u5927\u7684\u6587\u672c\u5206\u7c7b\u5668\u3002<\/p>\n \u8fd9\u91cc\u4ecb\u7ecd\u7684\u9879\u76ee\uff0c\u662fYoon Kim\u8bba\u6587\u300aConvolutional Neural Networks for Sentence Classification\uff08\u4f7f\u7528\u5377\u79ef\u795e\u7ecf\u7f51\u7edc\u8fdb\u884c\u53e5\u5b50\u5206\u7c7b\uff09\u300b\u7684\u7b80\u5355\u5b9e\u73b0\u3002\u8fd9\u4e2a\u8bba\u6587\u7684\u5730\u5740\u5982\u4e0b\uff1a<\/p>\n https:\/\/arxiv.org\/abs\/1408.5882<\/p>\n \u901a\u8fc7\u4e00\u4e2a\u7b80\u5355\u7684CNN\u5377\u79ef\u795e\u7ecf\u7f51\u7edc\uff0c\u53ea\u8fdb\u884c\u5f88\u5c11\u7684\u8d85\u53c2\u6570\u8c03\u6574\u548c\u9759\u6001\u77e2\u91cf\uff0c\u5c31\u53ef\u4ee5\u5f97\u5230\u51fa\u8272\u7684\u53e5\u5b50\u5206\u7c7b\u7ed3\u679c\u3002<\/p>\n \u6240\u9700\u8f6f\u4ef6\u73af\u5883\u5982\u4e0b\uff1a<\/p>\n Python 3<\/p>\n Tensorflow > 0.12<\/p>\n Numpy<\/p>\n GitHub\u5730\u5740\u5982\u4e0b\uff1a<\/p>\n https:\/\/github.com\/dennybritz\/cnn-text-classification-tf<\/p>\n \u9879\u76ee\u56db\uff1a\u56fe\u50cf\u5206\u7c7b\/\u7269\u4f53\u8bc6\u522b<\/p>\n \u56fe\u50cf\u5206\u7c7b\uff0c\u4e5f\u5c31\u662f\u8bad\u7ec3\u7cfb\u7edf\u8bc6\u522b\u732b\u732b\u72d7\u72d7\uff0c\u6216\u8005\u8f66\u9053\u3001\u6d77\u6ee9\u3001\u5929\u9645\u7ebf\u7b49\u573a\u666f\u3002\u8ba1\u7b97\u673a\u89c6\u89c9\u662f\u4e00\u4e2a\u8303\u56f4\u5de8\u5927\u7684\u9886\u57df\uff0c\u4ece\u9762\u90e8\u8bc6\u522b\u5230\u60c5\u611f\u8bc6\u522b\uff0c\u751a\u81f3\u53ef\u4ee5\u8fdb\u884c\u89c6\u89c9\u6c14\u4f53\u6cc4\u6f0f\u68c0\u6d4b\u3002\u867d\u7136\u5b9e\u73b0\u6d41\u7a0b\u5404\u5f02\uff0c\u4f46\u5e95\u5c42\u7cfb\u7edf\u662f\u76f8\u901a\u7684\u3002<\/p>\n \u6240\u4ee5\u4f5c\u8005\u521b\u5efa\u4e86\u4e00\u4e2aTensorFlow GitHub\u5e93\uff0c\u5176\u4e2d\u5305\u62ec\u4e24\u4e2a\u76f8\u4e92\u5173\u8054\u7684\u90e8\u5206\u3002\u7b2c\u4e00\u90e8\u5206\uff1a\u80fd\u591f\u8bc6\u522b1000\u4e2a\u5bf9\u8c61\u7684TensorFlow\u56fe\u50cf\u5206\u7c7b\u5668\u3002\u7b2c\u4e8c\u90e8\u5206\uff1a\u521b\u5efa\u5e76\u8bad\u7ec3\u4e00\u4e2a\u5168\u65b0\u7684\u56fe\u50cf\u5206\u7c7b\u5668\u3002<\/p>\n \u6211\u4eec\u5148\u6765\u770b\u770b\u7b2c\u4e00\u90e8\u5206\u3002<\/p>\n Google\u7684TensorFlow\u56fe\u50cf\u8bc6\u522b\u7cfb\u7edf\u662f\u76ee\u524d\u6700\u51c6\u786e\u7684\u56fe\u50cf\u5206\u7c7b\u8f6f\u4ef6\u3002\u6240\u8c13\u56fe\u50cf\u8bc6\u522b\uff0c\u5c31\u662f\u5bf9\u56fe\u7247\u4e2d\u7684\u5185\u5bb9\u8fdb\u884c\u8bc6\u522b\uff0c\u7136\u800c\u8fd9\u5e76\u975e\u5bf9\u4efb\u610f\u56fe\u7247\u90fd\u80fd\u8bc6\u522b\u3002<\/p>\n \u53ea\u6709\u88ab\u8bad\u7ec3\u8fc7\u7684\u5bf9\u8c61\uff0c\u7cfb\u7edf\u624d\u80fd\u8bc6\u522b\u3002\u4f8b\u5982\uff0c\u6211\u4eec\u7528\u4e09\u4e2a\u7c7b\u578b\u8bad\u7ec3\u5206\u7c7b\u5668\uff1a\u732b\u3001\u72d7\u548c\u725b\u3002\u4e09\u4e2a\u5206\u7c7b\u5668\u53ea\u80fd\u8bc6\u522b\u76f8\u5e94\u7c7b\u522b\u4e2d\u7684\u4e00\u4e2a\u3002\u5982\u679c\u7ed9\u51fa\u4e00\u5f20\u9a86\u9a7c\u7684\u56fe\u7247\u4f1a\u600e\u6837\uff1f\u56fe\u7247\u4ecd\u4f1a\u901a\u8fc7\u4e00\u4e2a\u5206\u7c7b\u5668\uff0c\u4f46\u662f\u7f6e\u4fe1\u7387\u4f1a\u975e\u5e38\u7684\u4f4e\u3002<\/p>\n \u5982\u4f55\u5feb\u901f\u521b\u5efa\u4e00\u4e2aTensorFlow\u56fe\u50cf\u5206\u7c7b\u5668\uff1f\u53ea\u9700\u8981\u7b80\u5355\u7684\u4e09\u6b65\u3002\u5f53\u7136\u524d\u63d0\u662f\u4f60\u5df2\u7ecf\u88c5\u597d\u4e86TensorFlow\uff0c\u800c\u4e14\u61c2\u7f16\u7a0b\u3001\u4f1a\u7528Python\u3002<\/p>\n \u7b2c\u4e00\u6b65\uff1a\u4e0b\u8f7d<\/p>\n \u4e0b\u8f7d\u9884\u8bad\u7ec3\u7684\u6a21\u578b\u3001\u56fe\u7247\u548c\u811a\u672c\u3002\u4f7f\u7528\u5982\u4e0b\u547d\u4ee4\u5373\u53ef\u3002<\/p>\n git clone https:\/\/github.com\/akshaypai\/tfClassifiercd tfClassifier<\/p>\n \u7b2c\u4e8c\u6b65\uff1a\u8fd0\u884c\u811a\u672c\u627e\u5230\u6700\u4f73\u9884\u6d4b<\/p>\n \u7ed9\u5b9a\u4e00\u4e2a\u56fe\u7247\uff0c\u8fd0\u884c\u811a\u672c\u6765\u8fdb\u884c\u5206\u7c7b\u3002\u9ed8\u8ba4\u60c5\u51b5\u4e0b\uff0c\u4f1a\u663e\u793a\u6700\u4f73\u9884\u6d4b\u3002<\/p>\n Python classifier.py –image_file file_path_to_image<\/p>\n \u4e3a\u4e86\u83b7\u5f97top n\u5206\u7c7b\uff0c\u53ef\u4ee5\u4f7f\u7528\u5982\u4e0b\u53c2\u6570\u3002<\/p>\n Python classifier.py –image_file file_path_to_image –num_top_predictions number_of_top_results<\/p>\n \u793a\u4f8b\uff1a\u6bd4\u65b9\u6211\u4eec\u7ed9\u51fa\u8fd9\u6837\u4e00\u5f20\u56fe\u7247\u3002<\/p>\n python classifier.py –image_file ~\/Pictures\/fruit.jpgpomegranate (score = 0.98216)<\/p>\n \u5206\u7c7b\u5668\u8868\u793a\uff1a\u8fd9\u662f\u77f3\u69b4\uff0c\u51c6\u786e\u738798%\u3002<\/p>\n \u7b2c\u4e09\u6b65\uff1a\u8fd0\u884c\u811a\u672c\u83b7\u5f97top n\u8bc6\u522b\u5206\u7c7b<\/p>\n \u73b0\u5728\u6211\u4eec\u5c1d\u8bd5\u7ed9\u51fa\u4e00\u4e2a\u5177\u6709\u66f4\u591a\u5c5e\u6027\u7684\u56fe\u50cf\uff0c\u5982\u4e0b\u56fe\u7684\u623f\u5b50\u3002<\/p>\n python classifier.py –image_file ~\/Pictures\/house.jpg –num_top_predictions 5picket fence, paling (score = 0.95750)worm fence, snake fence, snake-rail fence, Virginia fence (score = 0.03615)beacon, lighthouse, beacon light, pharos (score = 0.00018)boathouse (score = 0.00013)patio, terrace (score = 0.00007)<\/p>\n \u4ece\u4e0a\u9762\u7684\u7ed3\u679c\u53ef\u4ee5\u770b\u51fa\uff0c\u5206\u7c7b\u5668\u4ee595%\u7684\u53ef\u80fd\u6027\u9884\u6d4b\u56fe\u7247\u4e2d\u6709\u4e00\u4e2a\u6805\u680f\uff0c\u53e6\u5916\u5206\u7c7b\u5668\u8fd8\u53d1\u73b0\u4e86\u5176\u4ed6\u56f4\u680f\u3001\u5ead\u9662\u3001\u9732\u53f0\u7b49\u3002<\/p>\n \u5230\u6b64\u4e3a\u6b62\uff0c\u4f60\u5df2\u7ecf\u5b66\u5230\u5982\u4f55\u8bbe\u7f6eTensorFlow\u56fe\u50cf\u8bc6\u522b\u7cfb\u7edf\u3002\u867d\u7136\uff0c\u8fd9\u4e2a\u7cfb\u7edf\u88ab\u9650\u5236\u5728\u9884\u8bad\u7ec3\u6a21\u578b\u7684\u51e0\u4e2a\u5206\u7c7b\u5668\u4e4b\u95f4\u3002<\/p>\n \u90a3\u4e48\u6211\u4eec\u63a5\u4e0b\u6765\u770b\u770b\u7b2c\u4e8c\u90e8\u5206\uff0c\u5982\u4f55\u589e\u52a0\u4e00\u4e9b\u65b0\u7684\u5206\u7c7b\u5668\u3002<\/p>\n \u4e5f\u662f\u4e09\u4e2a\u6b65\u9aa4\u3002<\/p>\n \u7b2c\u4e00\u6b65\uff1a\u4e0b\u8f7d\u9884\u8bad\u7ec3\u6a21\u578b\u4ee5\u53ca\u6240\u9700\u811a\u672c<\/p>\n \u6211\u5df2\u7ecf\u628a\u5168\u90e8\u6240\u9700\u7684\u6587\u4ef6\u6574\u5408\u8fdb\u4e00\u4e2agit\u4ed3\u5e93\u91cc\u3002\u4f7f\u7528\u4e0b\u9762\u7684\u547d\u4ee4\u53ef\u4ee5\u4e0b\u8f7d\u3002<\/p>\n git clone https:\/\/github.com\/akshaypai\/tfClassifiercd tfClassifier<\/p>\n \u7b2c\u4e8c\u6b65\uff1a\u8bbe\u7f6e\u56fe\u50cf\u6587\u4ef6\u5939<\/p>\n \u8fd9\u4e2a\u6b65\u9aa4\u7528\u4e8e\u8bbe\u7f6e\u6587\u4ef6\u5939\u7ed3\u6784\uff0c\u4ee5\u4fbf\u6570\u636e\u6d41\u56fe\u53ef\u4ee5\u7b80\u5355\u5730\u62fe\u53d6\u5206\u7c7b\u3002\u5047\u8bbe\uff0c\u4f60\u60f3\u91cd\u65b0\u8bad\u7ec3\u4e94\u79cd\u65b0\u7684\u82b1\u6735\u5206\u7c7b\u5668\uff1a\u73ab\u7470\u3001\u90c1\u91d1\u9999\u3001\u84b2\u516c\u82f1\u3001\u4e94\u6708\u82b1\u548c\u4e07\u5bff\u83ca\uff0c\u90a3\u4e48\u9700\u8981\u5982\u4e0b\u7684\u4e09\u4e2a\u6b65\u9aa4\u6765\u521b\u5efa\u76f8\u5e94\u7684\u6587\u4ef6\u5939\u7ed3\u6784\uff1a<\/p>\n 1\u3001\u4e3a\u6bcf\u79cd\u82b1\u578b\u521b\u5efa\u4e00\u4e2a\u6587\u4ef6\u5939\uff0c\u6587\u4ef6\u5939\u7684\u540d\u79f0\u5c31\u662f\u7c7b\u578b\u7684\u540d\u79f0<\/p>\n 2\u3001\u5c06\u82b1\u7684\u6240\u6709\u56fe\u50cf\u6dfb\u52a0\u5230\u5404\u81ea\u7684\u6587\u4ef6\u5939\u4e2d\uff0c\u6240\u6709\u7684\u73ab\u7470\u653e\u5165\u73ab\u7470\u82b1\u6587\u4ef6\u5939<\/p>\n 3\u3001\u5c06\u6240\u6709\u7684\u6587\u4ef6\u5939\uff0c\u6dfb\u52a0\u5230\u4e00\u4e2a\u7236\u6587\u4ef6\u5939\u4e2d\uff0c\u53ef\u4ee5\u547d\u540d\u4e3a\uff1a\u82b1<\/p>\n \u7136\u540e\u6211\u4eec\u5c31\u5f97\u5230\u5982\u4e0b\u7684\u7ed3\u6784\uff1a<\/p>\n ~\/flowers~\/flowers\/roses\/img1.jpg~\/flowers\/roses\/img2.jpg…~\/flowers\/tulips\/tulips_img1.jpg~\/flowers\/tulips\/tulips_img2.jpg~\/flowers\/tulips\/tulips_img3.jpg…<\/p>\n \u8fd9\u6837\uff0c\u6587\u4ef6\u5939\u7ed3\u6784\u5df2\u7ecfOK\u4e86\u3002<\/p>\n \u7b2c\u4e09\u6b65\uff1a\u8fd0\u884c\u9884\u8bad\u7ec3\u811a\u672c<\/p>\n \u4f7f\u7528\u5982\u4e0b\u547d\u4ee4\u8fd0\u884c\u811a\u672c\u3002<\/p>\n python retrain.py –model_dir .\/inception –image_dir ~\/flowers –output_graph .\/output –how_many_training_steps 500<\/p>\n \u90e8\u5206\u547d\u4ee4\u884c\u53c2\u6570\uff1a<\/p>\n -model_dir\u8fd9\u4e2a\u53c2\u6570\u7ed9\u51fa\u4e86\u9884\u8bad\u7ec3\u6a21\u578b\u7684\u4f4d\u7f6e\u3002<\/p>\n -image_dir\u5728\u6b65\u9aa4\u4e8c\u4e2d\u521b\u5efa\u7684\u6587\u4ef6\u5939\u8def\u5f84<\/p>\n -output_graph\u5b58\u50a8\u65b0\u8bad\u7ec3\u56fe\u7684\u4f4d\u7f6e<\/p>\n -how_many_training_steps\u8fd9\u4ee3\u8868\u8981\u6267\u884c\u7684\u8fed\u4ee3\u6b21\u6570\uff0c\u9ed8\u8ba4\u4e3a4000<\/p>\n \u597d\u4e86\uff0c\u4ee5\u4e0a\u5c31\u662f\u5982\u4f55\u91cd\u65b0\u8bad\u7ec3\u4e00\u4e2aTensorFlow Inception\u6a21\u578b\u3002\u4e00\u65e6\u4f60\u6709\u4e86\u6a21\u578b\uff0c\u5c31\u80fd\u5f00\u59cb\u7528\u6765\u8fdb\u884c\u5206\u7c7b\u5904\u7406\u3002<\/p>\n \u7ed3\u8bba<\/p>\n \u6211\u5e0c\u671b\u4f60\u4eec\u4e5f\u80fd\u4f53\u4f1a\u5230\u4e0a\u8ff0\u8fd9\u4e9b\u9879\u76ee\u7684\u7cbe\u5f69\u4e4b\u5904\u3002\u5b66\u4e60\u4e00\u4ef6\u4e8b\u6700\u597d\u7684\u65b9\u6cd5\u5c31\u662f\u52a8\u624b\u53bb\u505a\u3002\u5982\u679c\u4f60\u8fd8\u6709\u66f4\u503c\u5f97\u63a8\u8350\u7684TensorFlow GitHub\u9879\u76ee\uff0c\u53ef\u4ee5\u7559\u8a00\u7ed9\u6211\u4eec\uff0c\u8ba9\u66f4\u591a\u7684\u4eba\u770b\u5230\u3002\u4eca\u5929\u5c31\u5148\u5230\u8fd9\u91cc\u5427\u3002\u4f11\u606f\uff0c\u4f11\u606f\u4e00\u4e0b~<\/p>\n <\/p>\n link:\u00a0http:\/\/www.techweb.com.cn\/news\/2017-07-31\/2566452.shtml<\/a><\/p>\n<\/a><\/p>\n<\/a><\/p>\n<\/a><\/p>\n<\/a><\/p>\n<\/a><\/p>\n<\/a><\/p>\n<\/a><\/p>\n<\/a><\/p>\n<\/a><\/p>\n