Question

我为每个手势（5 个手势）训练数据 700 张图像，验证测试数据 200 张图像和测试数据 150 张图像。我的模型是：

def get_model():
"""
Returns a compiled convolutional neural network model. Assume that the
`input_shape` of the first layer is `(IMG_WIDTH, IMG_HEIGHT, 3)`.
The output layer should have `NUM_CATEGORIES` units, one for each category.
"""
# Create a convolutional neural network
model = tf.keras.models.Sequential(
    [
    # Convolutional layer. Learn 32 filters using a 3x3 kernel
    tf.keras.layers.Conv2D(
        32, (5, 5), activation='relu', input_shape=(IMG_WIDTH, IMG_HEIGHT, 3)
    ),
    # Max-pooling layer, using 2x2 pool size
    tf.keras.layers.MaxPool2D(pool_size=(2, 2)),
    tf.keras.layers.Conv2D(
        64, (3, 3), activation='relu', input_shape=(IMG_WIDTH, IMG_HEIGHT, 3)
    ),
    # Max-pooling layer, using 2x2 pool size
    tf.keras.layers.MaxPool2D(pool_size=(2, 2)),
    tf.keras.layers.Conv2D(
        128, (3, 3), activation='relu', input_shape=((IMG_WIDTH), (IMG_HEIGHT), 3)
    ),
    tf.keras.layers.MaxPool2D(pool_size=(2, 2)),
    tf.keras.layers.Conv2D(
        256, (3, 3), activation='relu', input_shape=((IMG_WIDTH), (IMG_HEIGHT), 3)
    ),
    tf.keras.layers.MaxPool2D(pool_size=(2, 2)),
    
    tf.keras.layers.Flatten(),
    # Add a hidden layer with dropout
    tf.keras.layers.Dense(128, activation='relu'),
    tf.keras.layers.Dropout(0.3),
    # Add an output layer with output units for all 6 gestures
    tf.keras.layers.Dense(NUM_CATEGORIES, activation='softmax')
])
  
optimizer = tf.keras.optimizers.Adam(learning_rate=0.001)
model.compile(
    optimizer=optimizer ,
    loss="categorical_crossentropy",
    metrics=["accuracy"]
)
return model

模型拟合部分：

y_test = tf.keras.utils.to_categorical(labels_test)
y_train = tf.keras.utils.to_categorical(labels_train)
y_valid = tf.keras.utils.to_categorical(valid_label)

x_train = np.array(images_train)/255
x_test = np.array(images_test)/255

x_valid = np.array(valid_image)/255
# Get a compiled neural network
fitting_time = datetime.now()
model = get_model()

# Fit model on training data
model.fit(x_train, y_train, epochs=EPOCHS, validation_data = (x_valid, y_valid))
# Evaluate neural network performance
model.evaluate(x_test, y_test, verbose=2)

我多次更改学习率，但不起作用。它在火车数据中过度拟合。

如何避免过拟合以及为什么测试准确率如此低？

Answer 1

这种情况可能会发生，因为您在所有卷积层中都设置了 input_shape 参数。您应该只在第一个中定义它。

您的代码应如下所示：

def get_model():
"""
Returns a compiled convolutional neural network model. Assume that the
`input_shape` of the first layer is `(IMG_WIDTH, IMG_HEIGHT, 3)`.
The output layer should have `NUM_CATEGORIES` units, one for each category.
"""
# Create a convolutional neural network
model = tf.keras.models.Sequential(
    [
    # Convolutional layer. Learn 32 filters using a 3x3 kernel
    tf.keras.layers.Conv2D(
        32, (5, 5), activation='relu', input_shape=(IMG_WIDTH, IMG_HEIGHT, 3)
    ),
    # Max-pooling layer, using 2x2 pool size
    tf.keras.layers.MaxPool2D(pool_size=(2, 2)),
    tf.keras.layers.Conv2D(
        64, (3, 3), activation='relu'
    ),
    # Max-pooling layer, using 2x2 pool size
    tf.keras.layers.MaxPool2D(pool_size=(2, 2)),
    tf.keras.layers.Conv2D(
        128, (3, 3), activation='relu'
    ),
    tf.keras.layers.MaxPool2D(pool_size=(2, 2)),
    tf.keras.layers.Conv2D(
        256, (3, 3), activation='relu'
    ),
    tf.keras.layers.MaxPool2D(pool_size=(2, 2)),
    
    tf.keras.layers.Flatten(),
    # Add a hidden layer with dropout
    tf.keras.layers.Dense(128, activation='relu'),
    tf.keras.layers.Dropout(0.3),
    # Add an output layer with output units for all 6 gestures
    tf.keras.layers.Dense(NUM_CATEGORIES, activation='softmax')
])
  
optimizer = tf.keras.optimizers.Adam(learning_rate=0.001)
model.compile(
    optimizer=optimizer ,
    loss="categorical_crossentropy",
    metrics=["accuracy"]
)
return model

此外，您还应该检查训练和验证集中的分布是否相同。

如何避免训练数据中的过度拟合？

1 个答案: