卷积神经网络

main.py

@@ -1,61 +1,36 @@
-import os
-
 import tensorflow as tf
+config = tf.compat.v1.ConfigProto(gpu_options=tf.compat.v1.GPUOptions(allow_growth=True))
+sess = tf.compat.v1.Session(config=config)
 from tensorflow import keras
-import matplotlib.pyplot as plt
+fashion_mnist = keras.datasets.fashion_mnist
-
+(train_images, train_labels), (test_images, test_labels)  = fashion_mnist.load_data()
-print(tf.version.VERSION)
 
-(train_images, train_labels), (test_images, test_labels) = tf.keras.datasets.mnist.load_data()
+print(train_images.shape)
 
-plt.imshow(train_images[0])
+import tensorflow as tf
+from tensorflow import keras
+num_mnist = keras.datasets.mnist
+(train_images, train_labels), (test_images, test_labels) = num_mnist.load_data()
 
+train_images = train_images[:1000]
 train_labels = train_labels[:1000]
-test_labels = test_labels[:1000]
+test_images = train_images[:1000]
-
+test_labels = train_images[:1000]
-train_images = train_images[:1000].reshape(-1, 28 * 28) / 255.0
-test_images = test_images[:1000].reshape(-1, 28 * 28) / 255.0
-
-# 定义一个简单的序列模型
-def create_model():
-  model = tf.keras.models.Sequential([
-    keras.layers.Dense(512, activation='relu', input_shape=(784,)),
-    keras.layers.Dropout(0.2),
-    keras.layers.Dense(10)
-  ])
-
-  model.compile(optimizer='adam',
-                loss=tf.losses.SparseCategoricalCrossentropy(from_logits=True),
-                metrics=['accuracy'])
-
-  return model
-
-# 创建一个基本的模型实例
-model = create_model()
-
-# 显示模型的结构
-model.summary()
-
-checkpoint_path = "training_1/cp.ckpt"
-checkpoint_dir = os.path.dirname(checkpoint_path)
 
-# 创建一个保存模型权重的回调
+model = keras.Sequential()
-cp_callback = tf.keras.callbacks.ModelCheckpoint(filepath=checkpoint_path,
+model.add(keras.layers.Conv2D(8, (3,3), activation = 'relu', input_shape = (28,28,1)))
-                                                 save_weights_only=True,
+model.add(keras.layers.MaxPooling2D(2,2))
-                                                 verbose=1)
+model.add(keras.layers.Conv2D(8, (3,3), activation = 'relu'))
+model.add(keras.layers.MaxPooling2D(2,2))
 
-# 使用新的回调训练模型
+model.add(keras.layers.Flatten())
-model.fit(train_images,
+model.add(keras.layers.Dense(128, activation = tf.nn.relu))
-            train_labels,
+model.add(keras.layers.Dense(36, activation = tf.nn.softmax))
-            epochs=10,
-            batch_size=8,
-            validation_data=(test_images,test_labels),
-            callbacks=[cp_callback])  # 通过回调训练
 
-# 这可能会生成与保存优化程序状态相关的警告。
+train_images_scaled = train_images/255
-# 这些警告（以及整个笔记本中的类似警告）
+model.compile(optimizer = 'adam', loss = tf.losses.sparse_categorical_crossentropy, metrics = ['accuracy'])
-# 是防止过时使用，可以忽略。
 
-results = model.evaluate(test_images,  test_labels, verbose=2)
+history = model.fit(train_images_scaled.reshape(-1, 28, 28 ,1), train_labels, epochs = 10, batch_size=8)
 
-print(results)
+#test_images_scaled = test_images/255
+#results = model.evaluate(test_images_scaled.reshape(-1, 28, 28 ,1), test_labels)