a.初探神經網路:第一支神經網路

# a.初探神經網路:第一支神經網路 ###### tags: `Deep Learning by Keras God` ## 1.簡介 * **匯入手寫辨識資料集(Mnist)** ```python=+ from keras.datasets import mnist (train_img,train_label),(test_img,test_label) = mnist.load_data() ``` > ```11493376/11490434 [==============================] - 0s 0us/step``` ## * **建立模型及輸入層、輸出層** ```python=+ from keras.models import Sequential from keras.layers import Dense model = Sequential() model.add(Dense(512,activation = 'relu',input_shape = (28*28,))) model.add(Dense(10,activation = 'softmax')) ``` ## * **編譯模型** ```python=+ model.compile(optimizer = 'rmsprop', loss = 'categorical_crossentropy', metrics = ['accuracy']) ``` ## * **資料預處理 - 將圖片正規化** ```python=+ train_img = train_img.reshape(60000,28*28) train_img = train_img.astype('float32') / 255 test_img = test_img.reshape(10000,28*28) test_img = test_img.astype('float32') / 255 ``` ## * **資料預處理 - 將標籤(數字)轉為one-hot** ```python=+ from keras.utils import to_categorical train_label = to_categorical(train_label) test_label = to_categorical(test_label) ``` ## * **訓練模型** ```python=+ model.fit(train_img,train_label,epochs = 5,batch_size = 128) ``` > ```Epoch 1/5``` > ```469/469 [==============================] - 1s 3ms/step - loss: 0.2562 - accuracy: 0.9262``` > ```Epoch 2/5``` > ```469/469 [==============================] - 1s 3ms/step - loss: 0.1040 - accuracy: 0.9688``` > ```...``` > ```<tensorflow.python.keras.callbacks.History at 0x7fa1e0141a58>``` ## * **評估模型** ```python=+ test_loss,test_acc = model.evaluate(test_img,test_label) print('test_acc:',test_acc) ``` > ```test_acc: 0.9815999865531921``` ## 時間戳記 > [name=ZEOxO][time=Sun, Dec 06, 2020 15:30 PM][color=#907bf7]