Update mnist_dense.py

code/mnist_dense.py

@@ -4,21 +4,31 @@ from matplotlib import pyplot as plt
 from PIL import Image
 import numpy as np
 
-mnist = tf.keras.datasets.mnist
+# 数据准备
+mnist = tf.keras.datasets.mnist# 导入MNIST数据集
 
+# 指定要投入网络的训练集和测试集
 (x_train, y_train), (x_test, y_test) = mnist.load_data()
 x_train, x_test = x_train/255.0, x_test/255.0
 
 def creat_model():
     model = tf.keras.models.Sequential([
-        tf.keras.layers.Flatten(),
+        tf.keras.layers.Flatten(),  # 将输入层的数据压成一维数据
-        tf.keras.layers.Dense(128, activation='relu'),
+        # 全连接层
-        tf.keras.layers.Dense(10, activation='softmax')
+        # tf.keras.layers.Densen(神经元个数,activation='激活函数')
+        tf.keras.layers.Dense(128, activation='relu'),  # 128个神经元，激活函数为'relu'
+        tf.keras.layers.Dense(10, activation='softmax')  # 10个神经元，激活函数为'softmax'
     ])
 
-    model.compile(optimizer='adam',
+    # 使用compile()配置神经网络训练方法
+    # model.compile(optimizer=优化器,
+    #                     loss=损失函数,
+    #                     metrics=["评测指标"]
+    #                    )
+    model.compile(optimizer='adam',  # 优化器为'adam'
                   loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=False),
-                  metrics=["sparse_categorical_accuracy"])
+                  # 损失函数为'sparse_categorical_crossentropy'
+                  metrics=["sparse_categorical_accuracy"])  # y_以数值形式给出，y以独热码形式给出
     return model
 def model_fit(model,check_save_path):
 
@@ -26,12 +36,28 @@ def model_fit(model,check_save_path):
         print("load modals...")
         model.load_weights(check_save_path)
 
-    cp_callback = tf.keras.callbacks.ModelCheckpoint(filepath=check_save_path,
+    # 使用callbacks.ModelCheckpoint()保存模型
-                                                    save_weights_only=True,
+    # tf.keras.callbacks.ModelCheckpoint(filepath=保存模型的路径,
-                                                    save_best_only=True)
+    #                                                save_weights_only=若设置为True则仅保存模型权重,设置为False则保存整个模型
+    #                                                save_best_only=若设置为True将只保存在验证集上性能最好的模型
+    #                                                )
+    cp_callback = tf.keras.callbacks.ModelCheckpoint(filepath=check_save_path,  # 保存路径
+                                                     save_weights_only=True,  # 只保存模型权重
+                                                     save_best_only=True)  # 只保存在验证集上性能最好的模型
 
-    history = model.fit(x_train, y_train, batch_size=32, epochs=5, validation_split=0.15, validation_freq=1, callbacks=[cp_callback])
+    # 模型训练
-    model.summary()
+    # 在fit()中执行训练
+    # model.fit(训练集的输入特征,训练集的标签,
+    #             batch_size=,epochs=,
+    #             validation_split=从训练集中划分多少比例给测试集,
+    #             validation_freq=多少次epoch测试一次
+    #            )
+    history = model.fit(x_train, y_train,
+                        batch_size=32, epochs=5,  # 每一批batch大小为32,迭代次数epochs为5
+                        validation_split=0.15,  # 从训练集中划分15%给验证集
+                        validation_freq=1,  # 测试的间隔次数为1
+                        callbacks=[cp_callback])  # 回调参数
+    model.summary()  # 打印网络结构和参数统计
 
     final_loss, final_acc = model.evaluate(x_test, y_test, verbose=2)
     print("Model accuracy: ", final_acc, ", model loss: ", final_loss)
@@ -56,7 +82,8 @@ def model_fit(model,check_save_path):
 
 def eva_acc(str, model):
     model.load_weights(str)
-    final_loss, final_acc = model.evaluate(x_test, y_test, verbose=2)
+    # 准确率测试
+    final_loss, final_acc = model.evaluate(x_test, y_test, verbose=2)  # 输入数据和标签,输出损失和精确度,verbose = 2输出一行记录
     print("Model accuracy: ", final_acc, ", model loss: ", final_loss)
 
 if __name__=="__main__":