{"id":990886,"date":"2024-12-27T08:23:56","date_gmt":"2024-12-27T00:23:56","guid":{"rendered":"https:\/\/docs.pingcode.com\/ask\/ask-ask\/990886.html"},"modified":"2024-12-27T08:23:58","modified_gmt":"2024-12-27T00:23:58","slug":"python%e5%a6%82%e4%bd%95%e4%bd%bf%e7%94%a8keras%e5%ba%93","status":"publish","type":"post","link":"https:\/\/docs.pingcode.com\/ask\/990886.html","title":{"rendered":"Python\u5982\u4f55\u4f7f\u7528keras\u5e93"},"content":{"rendered":"

\"Python\u5982\u4f55\u4f7f\u7528keras\u5e93\"<\/p>\n

Python\u4f7f\u7528Keras\u5e93\u53ef\u4ee5\u901a\u8fc7\u4ee5\u4e0b\u51e0\u4e2a\u6b65\u9aa4\u5b9e\u73b0\uff1a\u5b89\u88c5Keras\u3001\u5bfc\u5165Keras\u5e93\u3001\u5b9a\u4e49\u6a21\u578b\u3001\u7f16\u8bd1\u6a21\u578b\u3001\u8bad\u7ec3\u6a21\u578b\u3001\u8bc4\u4f30\u6a21\u578b\u3001\u8fdb\u884c\u9884\u6d4b\u3002<\/strong> \u5176\u4e2d\uff0c\u5b9a\u4e49\u6a21\u578b\u662f\u4e00\u4e2a\u5173\u952e\u6b65\u9aa4\uff0c\u6d89\u53ca\u5230\u9009\u62e9\u5408\u9002\u7684\u6a21\u578b\u7ed3\u6784\u548c\u53c2\u6570\u3002\u63a5\u4e0b\u6765\uff0c\u6211\u5c06\u8be6\u7ec6\u4ecb\u7ecd\u5982\u4f55\u5b9a\u4e49\u6a21\u578b\u3002<\/p>\n<\/p>\n

\u5b9a\u4e49\u6a21\u578b\u662f\u4f7f\u7528Keras\u5e93\u8fdb\u884c\u6df1\u5ea6\u5b66\u4e60\u7684\u6838\u5fc3\u90e8\u5206\u3002Keras\u63d0\u4f9b\u4e86\u4e24\u79cd\u4e3b\u8981\u7684\u65b9\u6cd5\u6765\u5b9a\u4e49\u6a21\u578b\uff1aSequential\u6a21\u578b\u548c\u51fd\u6570\u5f0fAPI\u3002Sequential\u6a21\u578b\u662f\u4e00\u79cd\u7ebf\u6027\u5806\u53e0\u7684\u6a21\u578b\uff0c\u9002\u7528\u4e8e\u5927\u591a\u6570\u7b80\u5355\u7684\u7f51\u7edc\u7ed3\u6784\uff1b\u51fd\u6570\u5f0fAPI\u5219\u63d0\u4f9b\u4e86\u66f4\u5927\u7684\u7075\u6d3b\u6027\uff0c\u5141\u8bb8\u6784\u5efa\u590d\u6742\u7684\u7f51\u7edc\u7ed3\u6784\uff0c\u5982\u591a\u8f93\u5165\u591a\u8f93\u51fa\u6a21\u578b\u3001\u5171\u4eab\u5c42\u6a21\u578b\u7b49\u3002\u5728\u5b9a\u4e49\u6a21\u578b\u65f6\uff0c\u9700\u8981\u6839\u636e\u5177\u4f53\u95ee\u9898\u9009\u62e9\u5408\u9002\u7684\u6a21\u578b\u7c7b\u578b\u548c\u5c42\u6570\uff0c\u5e76\u914d\u7f6e\u6bcf\u4e00\u5c42\u7684\u53c2\u6570\uff0c\u5982\u6fc0\u6d3b\u51fd\u6570\u3001\u8282\u70b9\u6570\u7b49\u3002<\/p>\n<\/p>\n

\u4e00\u3001\u5b89\u88c5Keras<\/h3>\n<\/p>\n

Keras\u662f\u4e00\u4e2a\u9ad8\u5c42\u795e\u7ecf\u7f51\u7edcAPI\uff0c\u8fd0\u884c\u5728TensorFlow\u4e4b\u4e0a\u3002\u56e0\u6b64\uff0c\u5728\u5b89\u88c5Keras\u4e4b\u524d\uff0c\u9996\u5148\u9700\u8981\u786e\u4fdd\u5df2\u7ecf\u5b89\u88c5\u4e86TensorFlow\u3002\u53ef\u4ee5\u4f7f\u7528\u4ee5\u4e0b\u547d\u4ee4\u5b89\u88c5\uff1a<\/p>\n<\/p>\n

pip install tensorflow<\/p>\n
<\/code><\/pre>\n<\/p>\n
\u63a5\u7740\uff0c\u5b89\u88c5Keras\uff1a<\/p>\n<\/p>\n
pip install keras<\/p>\n
<\/code><\/pre>\n<\/p>\n
\u8fd9\u4e24\u4e2a\u5e93\u5b89\u88c5\u5b8c\u6210\u540e\uff0c\u6211\u4eec\u5c31\u53ef\u4ee5\u5728Python\u4e2d\u4f7f\u7528Keras\u4e86\u3002<\/p>\n<\/p>\n
\u4e8c\u3001\u5bfc\u5165Keras\u5e93<\/h3>\n<\/p>\n
\u5728\u5f00\u59cb\u5b9a\u4e49\u548c\u8bad\u7ec3\u6a21\u578b\u4e4b\u524d\uff0c\u6211\u4eec\u9700\u8981\u5bfc\u5165Keras\u5e93\u4e2d\u7684\u4e00\u4e9b\u6a21\u5757\u3002\u901a\u5e38\u6211\u4eec\u9700\u8981\u5bfc\u5165\u6a21\u578b\u7c7b\u548c\u4e00\u4e9b\u5e38\u7528\u7684\u5c42\u3002\u4ee5\u4e0b\u662f\u4e00\u4e2a\u5e38\u89c1\u7684\u5bfc\u5165\u8bed\u53e5\uff1a<\/p>\n<\/p>\n
from keras.models import Sequential<\/p>\n
from keras.layers import Dense, Activation<\/p>\n
<\/code><\/pre>\n<\/p>\n
\u5728\u8fd9\u4e2a\u4f8b\u5b50\u4e2d\uff0c\u6211\u4eec\u5bfc\u5165\u4e86Sequential<\/code>\u6a21\u578b\u7c7b\u548cDense<\/code>\u3001Activation<\/code>\u5c42\u3002<\/p>\n<\/p>\n
\u4e09\u3001\u5b9a\u4e49\u6a21\u578b<\/h3>\n<\/p>\n
1. \u4f7f\u7528Sequential\u6a21\u578b<\/h4>\n<\/p>\n
Sequential\u6a21\u578b\u662fKeras\u4e2d\u6700\u7b80\u5355\u7684\u6a21\u578b\uff0c\u9002\u5408\u4e8e\u7b80\u5355\u7684\u5c42\u6b21\u7ed3\u6784\u3002\u6211\u4eec\u53ef\u4ee5\u901a\u8fc7\u4ee5\u4e0b\u65b9\u5f0f\u5b9a\u4e49\u4e00\u4e2aSequential\u6a21\u578b\uff1a<\/p>\n<\/p>\n
model = Sequential()<\/p>\n
model.add(Dense(units=64, input_dim=100))<\/p>\n
model.add(Activation('relu'))<\/p>\n
model.add(Dense(units=10))<\/p>\n
model.add(Activation('softmax'))<\/p>\n
<\/code><\/pre>\n<\/p>\n
\u5728\u8fd9\u4e2a\u4f8b\u5b50\u4e2d\uff0c\u6211\u4eec\u5b9a\u4e49\u4e86\u4e00\u4e2a\u4e24\u5c42\u7684\u5168\u8fde\u63a5\u795e\u7ecf\u7f51\u7edc\u3002\u7b2c\u4e00\u5c42\u670964\u4e2a\u8282\u70b9\uff0c\u5e76\u4f7f\u7528ReLU\u6fc0\u6d3b\u51fd\u6570\u3002\u7b2c\u4e8c\u5c42\u670910\u4e2a\u8282\u70b9\uff0c\u4f7f\u7528Softmax\u6fc0\u6d3b\u51fd\u6570\uff0c\u9002\u5408\u4e8e\u591a\u5206\u7c7b\u95ee\u9898\u3002<\/p>\n<\/p>\n
2. \u4f7f\u7528\u51fd\u6570\u5f0fAPI<\/h4>\n<\/p>\n
\u51fd\u6570\u5f0fAPI\u63d0\u4f9b\u4e86\u66f4\u5927\u7684\u7075\u6d3b\u6027\uff0c\u9002\u5408\u4e8e\u590d\u6742\u7684\u7f51\u7edc\u7ed3\u6784\u3002\u4ee5\u4e0b\u662f\u4e00\u4e2a\u7b80\u5355\u7684\u4f8b\u5b50\uff1a<\/p>\n<\/p>\n
from keras.layers import Input<\/p>\n
from keras.models import Model<\/p>\n
inputs = Input(shape=(100,))<\/p>\n
x = Dense(64, activation='relu')(inputs)<\/p>\n
outputs = Dense(10, activation='softmax')(x)<\/p>\n
model = Model(inputs=inputs, outputs=outputs)<\/p>\n
<\/code><\/pre>\n<\/p>\n
\u5728\u8fd9\u4e2a\u4f8b\u5b50\u4e2d\uff0c\u6211\u4eec\u9996\u5148\u5b9a\u4e49\u4e86\u8f93\u5165\u5c42\uff0c\u7136\u540e\u9010\u5c42\u6784\u5efa\u7f51\u7edc\uff0c\u6700\u540e\u5c06\u8f93\u5165\u5c42\u548c\u8f93\u51fa\u5c42\u4f20\u9012\u7ed9Model<\/code>\u7c7b\u6765\u5b9a\u4e49\u6a21\u578b\u3002<\/p>\n<\/p>\n
\u56db\u3001\u7f16\u8bd1\u6a21\u578b<\/h3>\n<\/p>\n
\u5728\u5b9a\u4e49\u597d\u6a21\u578b\u7ed3\u6784\u540e\uff0c\u6211\u4eec\u9700\u8981\u7f16\u8bd1\u6a21\u578b\u3002\u7f16\u8bd1\u6a21\u578b\u65f6\u9700\u8981\u6307\u5b9a\u4f18\u5316\u5668\u3001\u635f\u5931\u51fd\u6570\u548c\u8bc4\u4f30\u6307\u6807\u3002\u4ee5\u4e0b\u662f\u4e00\u4e2a\u7f16\u8bd1\u6a21\u578b\u7684\u4f8b\u5b50\uff1a<\/p>\n<\/p>\n
model.compile(optimizer='adam',<\/p>\n
              loss='categorical_crossentropy',<\/p>\n
              metrics=['accuracy'])<\/p>\n
<\/code><\/pre>\n<\/p>\n
\u5728\u8fd9\u4e2a\u4f8b\u5b50\u4e2d\uff0c\u6211\u4eec\u9009\u62e9Adam\u4f18\u5316\u5668\u548c\u4ea4\u53c9\u71b5\u635f\u5931\u51fd\u6570\uff0c\u5e76\u6307\u5b9a\u4e86\u51c6\u786e\u7387\u4f5c\u4e3a\u8bc4\u4f30\u6307\u6807\u3002<\/p>\n<\/p>\n
\u4e94\u3001\u8bad\u7ec3\u6a21\u578b<\/h3>\n<\/p>\n
\u7f16\u8bd1\u597d\u6a21\u578b\u540e\uff0c\u6211\u4eec\u53ef\u4ee5\u4f7f\u7528\u8bad\u7ec3\u6570\u636e\u6765\u8bad\u7ec3\u6a21\u578b\u3002\u8bad\u7ec3\u6a21\u578b\u65f6\u9700\u8981\u6307\u5b9a\u8f93\u5165\u6570\u636e\u3001\u6807\u7b7e\u3001\u6279\u5927\u5c0f\u548c\u8bad\u7ec3\u7684\u8f6e\u6570\u3002\u4ee5\u4e0b\u662f\u4e00\u4e2a\u8bad\u7ec3\u6a21\u578b\u7684\u4f8b\u5b50\uff1a<\/p>\n<\/p>\n
model.fit(x_trAI<\/a>n, y_train, batch_size=32, epochs=10)<\/p>\n
<\/code><\/pre>\n<\/p>\n
\u5728\u8fd9\u4e2a\u4f8b\u5b50\u4e2d\uff0c\u6211\u4eec\u4f7f\u7528\u6279\u5927\u5c0f\u4e3a32\u7684\u6570\u636e\u8fdb\u884c10\u8f6e\u8bad\u7ec3\u3002<\/p>\n<\/p>\n
\u516d\u3001\u8bc4\u4f30\u6a21\u578b<\/h3>\n<\/p>\n
\u8bad\u7ec3\u5b8c\u6210\u540e\uff0c\u6211\u4eec\u53ef\u4ee5\u4f7f\u7528\u6d4b\u8bd5\u6570\u636e\u6765\u8bc4\u4f30\u6a21\u578b\u7684\u6027\u80fd\u3002\u4ee5\u4e0b\u662f\u4e00\u4e2a\u8bc4\u4f30\u6a21\u578b\u7684\u4f8b\u5b50\uff1a<\/p>\n<\/p>\n
loss, accuracy = model.evaluate(x_test, y_test)<\/p>\n
print(f"Test loss: {loss}")<\/p>\n
print(f"Test accuracy: {accuracy}")<\/p>\n
<\/code><\/pre>\n<\/p>\n
\u5728\u8fd9\u4e2a\u4f8b\u5b50\u4e2d\uff0c\u6211\u4eec\u8ba1\u7b97\u4e86\u6a21\u578b\u5728\u6d4b\u8bd5\u6570\u636e\u4e0a\u7684\u635f\u5931\u548c\u51c6\u786e\u7387\u3002<\/p>\n<\/p>\n
\u4e03\u3001\u8fdb\u884c\u9884\u6d4b<\/h3>\n<\/p>\n
\u6700\u540e\uff0c\u6211\u4eec\u53ef\u4ee5\u4f7f\u7528\u8bad\u7ec3\u597d\u7684\u6a21\u578b\u8fdb\u884c\u9884\u6d4b\u3002\u4ee5\u4e0b\u662f\u4e00\u4e2a\u9884\u6d4b\u7684\u4f8b\u5b50\uff1a<\/p>\n<\/p>\n
predictions = model.predict(x_new)<\/p>\n
<\/code><\/pre>\n<\/p>\n
\u5728\u8fd9\u4e2a\u4f8b\u5b50\u4e2d\uff0c\u6211\u4eec\u5bf9\u65b0\u6570\u636ex_new<\/code>\u8fdb\u884c\u4e86\u9884\u6d4b\uff0c\u8fd4\u56de\u7684\u7ed3\u679c\u662f\u6bcf\u4e2a\u7c7b\u7684\u6982\u7387\u3002<\/p>\n<\/p>\n
\u516b\u3001Keras\u4e2d\u7684\u5176\u4ed6\u529f\u80fd<\/h3>\n<\/p>\n
\u9664\u4e86\u57fa\u672c\u7684\u6a21\u578b\u5b9a\u4e49\u3001\u8bad\u7ec3\u548c\u8bc4\u4f30\u5916\uff0cKeras\u8fd8\u63d0\u4f9b\u4e86\u8bb8\u591a\u5176\u4ed6\u7684\u529f\u80fd\uff0c\u5982\u6a21\u578b\u4fdd\u5b58\u548c\u52a0\u8f7d\u3001\u56de\u8c03\u51fd\u6570\u3001\u6570\u636e\u9884\u5904\u7406\u7b49\u3002<\/p>\n<\/p>\n
1. \u6a21\u578b\u4fdd\u5b58\u548c\u52a0\u8f7d<\/h4>\n<\/p>\n
\u5728\u8bad\u7ec3\u5b8c\u6210\u540e\uff0c\u6211\u4eec\u53ef\u4ee5\u5c06\u6a21\u578b\u4fdd\u5b58\u5230\u6587\u4ef6\u4e2d\uff0c\u4ee5\u4fbf\u5728\u672a\u6765\u4f7f\u7528\u3002\u4ee5\u4e0b\u662f\u4fdd\u5b58\u548c\u52a0\u8f7d\u6a21\u578b\u7684\u4f8b\u5b50\uff1a<\/p>\n<\/p>\n
# \u4fdd\u5b58\u6a21\u578b<\/p>\n
model.save('my_model.h5')<\/p>\n
\u52a0\u8f7d\u6a21\u578b<\/strong><\/h2>\n
from keras.models import load_model<\/p>\n
model = load_model('my_model.h5')<\/p>\n
<\/code><\/pre>\n<\/p>\n
2. \u56de\u8c03\u51fd\u6570<\/h4>\n<\/p>\n
Keras\u63d0\u4f9b\u4e86\u4e00\u4e9b\u56de\u8c03\u51fd\u6570\uff0c\u7528\u4e8e\u5728\u8bad\u7ec3\u8fc7\u7a0b\u4e2d\u8fdb\u884c\u67d0\u4e9b\u64cd\u4f5c\uff0c\u5982\u8bb0\u5f55\u65e5\u5fd7\u3001\u4fdd\u5b58\u6a21\u578b\u3001\u8c03\u6574\u5b66\u4e60\u7387\u7b49\u3002\u4ee5\u4e0b\u662f\u4f7f\u7528\u56de\u8c03\u51fd\u6570\u7684\u4f8b\u5b50\uff1a<\/p>\n<\/p>\n
from keras.callbacks import EarlyStopping<\/p>\n
early_stopping = EarlyStopping(monitor='val_loss', patience=2)<\/p>\n
model.fit(x_train, y_train, validation_split=0.2, epochs=10, callbacks=[early_stopping])<\/p>\n
<\/code><\/pre>\n<\/p>\n
\u5728\u8fd9\u4e2a\u4f8b\u5b50\u4e2d\uff0c\u6211\u4eec\u4f7f\u7528\u4e86\u65e9\u505c\u6cd5\uff0c\u6839\u636e\u9a8c\u8bc1\u96c6\u7684\u635f\u5931\u6765\u51b3\u5b9a\u662f\u5426\u63d0\u524d\u505c\u6b62\u8bad\u7ec3\u3002<\/p>\n<\/p>\n
3. \u6570\u636e\u9884\u5904\u7406<\/h4>\n<\/p>\n
Keras\u63d0\u4f9b\u4e86\u4e00\u4e9b\u5de5\u5177\u6765\u5e2e\u52a9\u8fdb\u884c\u6570\u636e\u9884\u5904\u7406\uff0c\u5982\u56fe\u50cf\u589e\u5f3a\u3001\u5f52\u4e00\u5316\u7b49\u3002\u4ee5\u4e0b\u662f\u4e00\u4e2a\u7b80\u5355\u7684\u4f8b\u5b50\uff1a<\/p>\n<\/p>\n
from keras.preprocessing.image import ImageDataGenerator<\/p>\n
datagen = ImageDataGenerator(rotation_range=20, width_shift_range=0.2, height_shift_range=0.2, horizontal_flip=True)<\/p>\n
datagen.fit(x_train)<\/p>\n
model.fit(datagen.flow(x_train, y_train, batch_size=32), epochs=10)<\/p>\n
<\/code><\/pre>\n<\/p>\n
\u5728\u8fd9\u4e2a\u4f8b\u5b50\u4e2d\uff0c\u6211\u4eec\u4f7f\u7528ImageDataGenerator\u8fdb\u884c\u56fe\u50cf\u6570\u636e\u589e\u5f3a\u3002<\/p>\n<\/p>\n
\u4e5d\u3001\u603b\u7ed3<\/h3>\n<\/p>\n
Keras\u662f\u4e00\u4e2a\u529f\u80fd\u5f3a\u5927\u7684\u6df1\u5ea6\u5b66\u4e60\u5e93\uff0c\u9002\u5408\u4e8e\u5feb\u901f\u6784\u5efa\u548c\u6d4b\u8bd5\u795e\u7ecf\u7f51\u7edc\u6a21\u578b\u3002\u901a\u8fc7\u672c\u6587\u7684\u4ecb\u7ecd\uff0c\u60a8\u5e94\u8be5\u5bf9\u5982\u4f55\u5728Python\u4e2d\u4f7f\u7528Keras\u5e93\u6709\u4e86\u4e00\u4e2a\u521d\u6b65\u7684\u4e86\u89e3\u3002\u638c\u63e1\u4e86\u5b89\u88c5\u3001\u5bfc\u5165\u3001\u5b9a\u4e49\u3001\u7f16\u8bd1\u3001\u8bad\u7ec3\u3001\u8bc4\u4f30\u548c\u9884\u6d4b\u7b49\u57fa\u672c\u6b65\u9aa4\u540e\uff0c\u60a8\u53ef\u4ee5\u6839\u636e\u5177\u4f53\u9700\u6c42\u8c03\u6574\u6a21\u578b\u7684\u7ed3\u6784\u548c\u53c2\u6570\uff0c\u4f7f\u7528Keras\u63d0\u4f9b\u7684\u5176\u4ed6\u529f\u80fd\u6765\u5b9e\u73b0\u66f4\u590d\u6742\u7684\u6df1\u5ea6\u5b66\u4e60\u5e94\u7528\u3002<\/p>\n<\/p>\n
\u76f8\u5173\u95ee\u7b54FAQs\uff1a<\/strong><\/h2>\n
 \u5982\u4f55\u5b89\u88c5Keras\u5e93\u4ee5\u4fbf\u5728Python\u4e2d\u4f7f\u7528\uff1f<\/strong>
\u8981\u5728Python\u4e2d\u4f7f\u7528Keras\u5e93\uff0c\u9996\u5148\u9700\u8981\u786e\u4fdd\u4f60\u5df2\u7ecf\u5b89\u88c5\u4e86Python\u73af\u5883\u3002\u53ef\u4ee5\u4f7f\u7528pip\u547d\u4ee4\u5728\u7ec8\u7aef\u6216\u547d\u4ee4\u63d0\u793a\u7b26\u4e2d\u8f93\u5165\u4ee5\u4e0b\u547d\u4ee4\u8fdb\u884c\u5b89\u88c5\uff1a  <\/p>\n
pip install keras\n<\/code><\/pre>\n
\u5982\u679c\u4f60\u4f7f\u7528\u7684\u662fAnaconda\uff0c\u4e5f\u53ef\u4ee5\u901a\u8fc7Conda\u5b89\u88c5Keras\uff1a  <\/p>\n
conda install -c conda-forge keras\n<\/code><\/pre>\n
\u5b89\u88c5\u5b8c\u6210\u540e\uff0c\u4f60\u53ef\u4ee5\u901a\u8fc7\u5728Python\u4ee3\u7801\u4e2d\u5bfc\u5165Keras\u6765\u5f00\u59cb\u4f7f\u7528\u5b83\u3002<\/p>\n
Keras\u5e93\u7684\u4e3b\u8981\u529f\u80fd\u548c\u7279\u70b9\u662f\u4ec0\u4e48\uff1f<\/strong>
Keras\u662f\u4e00\u4e2a\u9ad8\u5c42\u6b21\u7684\u6df1\u5ea6\u5b66\u4e60API\uff0c\u5177\u6709\u7b80\u6d01\u548c\u6613\u7528\u7684\u7279\u70b9\u3002\u5b83\u652f\u6301\u591a\u79cd\u540e\u7aef\uff0c\u4f8b\u5982TensorFlow\u3001Theano\u548cCNTK\u3002Keras\u7684\u4e3b\u8981\u529f\u80fd\u5305\u62ec\uff1a  <\/p>\n
    \n
  • \u6784\u5efa\u548c\u8bad\u7ec3\u795e\u7ecf\u7f51\u7edc<\/strong>\uff1a\u53ef\u4ee5\u8f7b\u677e\u521b\u5efa\u4e0d\u540c\u7c7b\u578b\u7684\u795e\u7ecf\u7f51\u7edc\u7ed3\u6784\uff0c\u5982\u5377\u79ef\u795e\u7ecf\u7f51\u7edc\uff08CNN\uff09\u548c\u9012\u5f52\u795e\u7ecf\u7f51\u7edc\uff08RNN\uff09\u3002  <\/li>\n
  • \u6a21\u578b\u8bc4\u4f30\u548c\u9884\u6d4b<\/strong>\uff1a\u63d0\u4f9b\u4e86\u7b80\u5355\u7684\u65b9\u6cd5\u6765\u8bc4\u4f30\u6a21\u578b\u6027\u80fd\u548c\u8fdb\u884c\u9884\u6d4b\u3002  <\/li>\n
  • \u9884\u8bad\u7ec3\u6a21\u578b<\/strong>\uff1a\u5185\u7f6e\u4e86\u4e00\u4e9b\u6d41\u884c\u7684\u9884\u8bad\u7ec3\u6a21\u578b\uff0c\u9002\u5408\u8fc1\u79fb\u5b66\u4e60\u3002  <\/li>\n
  • \u652f\u6301\u591a\u79cd\u6570\u636e\u7c7b\u578b<\/strong>\uff1a\u53ef\u4ee5\u5904\u7406\u56fe\u50cf\u3001\u6587\u672c\u3001\u65f6\u95f4\u5e8f\u5217\u7b49\u591a\u79cd\u6570\u636e\u7c7b\u578b\u3002<\/li>\n<\/ul>\n
    \u6211\u8be5\u5982\u4f55\u5f00\u59cb\u4f7f\u7528Keras\u6784\u5efa\u7b2c\u4e00\u4e2a\u6df1\u5ea6\u5b66\u4e60\u6a21\u578b\uff1f<\/strong>
    \u6784\u5efa\u7b2c\u4e00\u4e2a\u6df1\u5ea6\u5b66\u4e60\u6a21\u578b\u7684\u6b65\u9aa4\u76f8\u5bf9\u7b80\u5355\u3002\u9996\u5148\uff0c\u5bfc\u5165\u5fc5\u8981\u7684\u5e93\u548c\u6a21\u5757\uff0c\u5305\u62ecKeras\u548c\u6240\u9700\u7684\u5c42\u3002\u63a5\u7740\uff0c\u5b9a\u4e49\u6a21\u578b\u67b6\u6784\uff0c\u4f8b\u5982\u9009\u62e9Sequential\u6a21\u578b\u5e76\u6dfb\u52a0\u5404\u5c42\u3002\u4ee5\u4e0b\u662f\u4e00\u4e2a\u7b80\u5355\u7684\u793a\u4f8b\uff1a  <\/p>\n
    from keras.models import Sequential\nfrom keras.layers import Dense\n\nmodel = Sequential()\nmodel.add(Dense(units=64, activation='relu', input_dim=10))\nmodel.add(Dense(units=1, activation='sigmoid'))\n\nmodel.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])\n<\/code><\/pre>\n
    \u5728\u6784\u5efa\u597d\u6a21\u578b\u540e\uff0c\u53ef\u4ee5\u4f7f\u7528fit<\/code>\u65b9\u6cd5\u6765\u8bad\u7ec3\u6a21\u578b\uff0c\u5e76\u901a\u8fc7\u63d0\u4f9b\u8bad\u7ec3\u6570\u636e\u548c\u6807\u7b7e\u6765\u8fdb\u884c\u8bad\u7ec3\u3002<\/p>\n","protected":false},"excerpt":{"rendered":"Python\u4f7f\u7528Keras\u5e93\u53ef\u4ee5\u901a\u8fc7\u4ee5\u4e0b\u51e0\u4e2a\u6b65\u9aa4\u5b9e\u73b0\uff1a\u5b89\u88c5Keras\u3001\u5bfc\u5165Keras\u5e93\u3001\u5b9a\u4e49\u6a21\u578b\u3001\u7f16\u8bd1\u6a21\u578b\u3001\u8bad […]","protected":false},"author":3,"featured_media":990898,"comment_status":"closed","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"_acf_changed":false,"footnotes":""},"categories":[37],"tags":[],"acf":[],"_links":{"self":[{"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/posts\/990886"}],"collection":[{"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/users\/3"}],"replies":[{"embeddable":true,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/comments?post=990886"}],"version-history":[{"count":"1","href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/posts\/990886\/revisions"}],"predecessor-version":[{"id":990902,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/posts\/990886\/revisions\/990902"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/media\/990898"}],"wp:attachment":[{"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/media?parent=990886"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/categories?post=990886"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/docs.pingcode.com\/wp-json\/wp\/v2\/tags?post=990886"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}