Alexnet的tensorflow实现

代码：https://github.com/kratzert/finetune_alexnet_with_tensorflow （juypter notebook）

Alexnet：

 代码解析：               

一、首先定义一个类：

class AlexNet(object):

二、之后定义类的初始化：

注意下，'bvlc_alexnet.npy'为别人训练好的Alexnet的权值，x为输入图像，keep_prob为全连接层dropout的概率，num_classes是分类数

def __init__(self, x, keep_prob, num_classes, skip_layer,
                 weights_path='DEFAULT'):
        """Create the graph of the AlexNet model.

        Args:
            x: Placeholder for the input tensor.
            keep_prob: Dropout probability.
            num_classes: Number of classes in the dataset.
            skip_layer: List of names of the layer, that get trained from
                scratch
            weights_path: Complete path to the pretrained weight file, if it
                isn't in the same folder as this code
        """
        # Parse input arguments into class variables
        self.X = x
        self.NUM_CLASSES = num_classes
        self.KEEP_PROB = keep_prob
        self.SKIP_LAYER = skip_layer

        if weights_path == 'DEFAULT':
            self.WEIGHTS_PATH = 'bvlc_alexnet.npy'
        else:
            self.WEIGHTS_PATH = weights_path

        # Call the create function to build the computational graph of AlexNet
        self.create()

 三、初始化函数调用了create():

def create(self):
        """Create the network graph."""
        # 1st Layer: Conv (w ReLu) -> Lrn -> Pool
        conv1 = conv(self.X, 11, 11, 96, 4, 4, padding='VALID', name='conv1')
        norm1 = lrn(conv1, 2, 2e-05, 0.75, name='norm1')
        pool1 = max_pool(norm1, 3, 3, 2, 2, padding='VALID', name='pool1')
        
        # 2nd Layer: Conv (w ReLu)  -> Lrn -> Pool with 2 groups
        conv2 = conv(pool1, 5, 5, 256, 1, 1, groups=2, name='conv2')
        norm2 = lrn(conv2, 2, 2e-05, 0.75, name='norm2')
        pool2 = max_pool(norm2, 3, 3, 2, 2, padding='VALID', name='pool2')
        
        # 3rd Layer: Conv (w ReLu)
        conv3 = conv(pool2, 3, 3, 384, 1, 1, name='conv3')

        # 4th Layer: Conv (w ReLu) splitted into two groups
        conv4 = conv(conv3, 3, 3, 384, 1, 1, groups=2, name='conv4')

        # 5th Layer: Conv (w ReLu) -> Pool splitted into two groups
        conv5 = conv(conv4, 3, 3, 256, 1, 1, groups=2, name='conv5')
        pool5 = max_pool(conv5, 3, 3, 2, 2, padding='VALID', name='pool5')

        # 6th Layer: Flatten -> FC (w ReLu) -> Dropout
        flattened = tf.reshape(pool5, [-1, 6*6*256])
        fc6 = fc(flattened, 6*6*256, 4096, name='fc6')
        dropout6 = dropout(fc6, self.KEEP_PROB)

        # 7th Layer: FC (w ReLu) -> Dropout
        fc7 = fc(dropout6, 4096, 4096, name='fc7')
        dropout7 = dropout(fc7, self.KEEP_PROB)

        # 8th Layer: FC and return unscaled activations
        self.fc8 = fc(dropout7, 4096, self.NUM_CLASSES, relu=False, name='fc8')

四、conv层：

注意，由于是用的别人训练好的权值，所以这里一样，仿照原作者的Alexnet（在两个GPU上训练）。

其中语法：

1、for i, k in zip(input_groups, weight_groups)

for (a,b,c) in zip(ta,tb,tc):从多个等长序列中取一个元素作为新序列

2、tf.split 

tf.split( value, num_or_size_splits, axis=0, num=None, name='split' )value：准备切分的张量 
num_or_size_splits：准备切成几份 
axis : 准备在第几个维度上进行切割 
其中分割方式分为两种 
1. 如果num_or_size_splits 传入的 是一个整数，那直接在axis=D这个维度上把张量平均切分成几个小张量 
2. 如果num_or_size_splits 传入的是一个向量（这里向量各个元素的和要跟原本这个维度的数值相等）就根据这个向量有几个元素分为几项）

3、tf.concat()是tensorflow中用来拼接张量的函数

def conv(x, filter_height, filter_width, num_filters, stride_y, stride_x, name,
         padding='SAME', groups=1):
    """Create a convolution layer.

    Adapted from: https://github.com/ethereon/caffe-tensorflow
    """
    # Get number of input channels
    input_channels = int(x.get_shape()[-1])

    # Create lambda function for the convolution
    convolve = lambda i, k: tf.nn.conv2d(i, k,
                                         strides=[1, stride_y, stride_x, 1],
                                         padding=padding)

    with tf.variable_scope(name) as scope:
        # Create tf variables for the weights and biases of the conv layer
        weights = tf.get_variable('weights', shape=[filter_height,
                                                    filter_width,
                                                    input_channels/groups,
                                                    num_filters])
        biases = tf.get_variable('biases', shape=[num_filters])

    if groups == 1:
        conv = convolve(x, weights)

    # In the cases of multiple groups, split inputs & weights and
    else:
        # Split input and weights and convolve them separately
        input_groups = tf.split(axis=3, num_or_size_splits=groups, value=x)
        weight_groups = tf.split(axis=3, num_or_size_splits=groups,
                                 value=weights)
        output_groups = [convolve(i, k) for i, k in zip(input_groups, weight_groups)]

        # Concat the convolved output together again
        conv = tf.concat(axis=3, values=output_groups)

    # Add biases
    bias = tf.reshape(tf.nn.bias_add(conv, biases), tf.shape(conv))

    # Apply relu function
    relu = tf.nn.relu(bias, name=scope.name)

    return relu

五、全连接层：

这里是xw+b，不是wx+b，注意下：

def fc(x, num_in, num_out, name, relu=True):
    """Create a fully connected layer."""
    with tf.variable_scope(name) as scope:

        # Create tf variables for the weights and biases
        weights = tf.get_variable('weights', shape=[num_in, num_out],
                                  trainable=True)
        biases = tf.get_variable('biases', [num_out], trainable=True)

        # Matrix multiply weights and inputs and add bias
        act = tf.nn.xw_plus_b(x, weights, biases, name=scope.name)

    if relu:
        # Apply ReLu non linearity
        relu = tf.nn.relu(act)
        return relu
    else:
        return act

六、池化层、dropout层和LRN层（Andrew Ng说LRN没什么用：）

def max_pool(x, filter_height, filter_width, stride_y, stride_x, name,
             padding='SAME'):
    """Create a max pooling layer."""
    return tf.nn.max_pool(x, ksize=[1, filter_height, filter_width, 1],
                          strides=[1, stride_y, stride_x, 1],
                          padding=padding, name=name)


def lrn(x, radius, alpha, beta, name, bias=1.0):
    """Create a local response normalization layer."""
    return tf.nn.local_response_normalization(x, depth_radius=radius,
                                              alpha=alpha, beta=beta,
                                              bias=bias, name=name)


def dropout(x, keep_prob):
    """Create a dropout layer."""
    return tf.nn.dropout(x, keep_prob)

七、加载别人训练好的权重 ：

def load_initial_weights(self, session):
        """Load weights from file into network.

        As the weights from http://www.cs.toronto.edu/~guerzhoy/tf_alexnet/
        come as a dict of lists (e.g. weights['conv1'] is a list) and not as
        dict of dicts (e.g. weights['conv1'] is a dict with keys 'weights' &
        'biases') we need a special load function
        """
        # Load the weights into memory
        weights_dict = np.load(self.WEIGHTS_PATH, encoding='bytes').item()

        # Loop over all layer names stored in the weights dict
        for op_name in weights_dict:

            # Check if layer should be trained from scratch
            if op_name not in self.SKIP_LAYER:

                with tf.variable_scope(op_name, reuse=True):

                    # Assign weights/biases to their corresponding tf variable
                    for data in weights_dict[op_name]:

                        # Biases
                        if len(data.shape) == 1:
                            var = tf.get_variable('biases', trainable=False)
                            session.run(var.assign(data))

                        # Weights
                        else:
                            var = tf.get_variable('weights', trainable=False)
                            session.run(var.assign(data))

八、现在已经定义好了Alexnet，可以调用模型

最后一层为softmax层，model = AlexNet(x, keep_prob, 1000, [])中的1000为最后一层的输出神经元数目，在这一层加上softmax即可。

from alexnet import AlexNet
from caffe_classes import class_names

#placeholder for input and dropout rate
x = tf.placeholder(tf.float32, [1, 227, 227, 3])
keep_prob = tf.placeholder(tf.float32)

#create model with default config ( == no skip_layer and 1000 units in the last layer)
model = AlexNet(x, keep_prob, 1000, [])

#define activation of last layer as score
score = model.fc8

#create op to calculate softmax 
softmax = tf.nn.softmax(score)

九、最后开始测试自己的图片：

注意将自己的图片reszie到227*227*3，因为这是Alexnet定义时，要求输入的大小

with tf.Session() as sess:
    
    # Initialize all variables
    sess.run(tf.global_variables_initializer())
    
    # Load the pretrained weights into the model
    model.load_initial_weights(sess)
    
    # Create figure handle
    fig2 = plt.figure(figsize=(15,6))
    
    # Loop over all images
    for i, image in enumerate(imgs):
        
        # Convert image to float32 and resize to (227x227)
        img = cv2.resize(image.astype(np.float32), (227,227))
        
        # Subtract the ImageNet mean
        img -= imagenet_mean
        
        # Reshape as needed to feed into model
        img = img.reshape((1,227,227,3))
        
        # Run the session and calculate the class probability
        probs = sess.run(softmax, feed_dict={x: img, keep_prob: 1})
        
        # Get the class name of the class with the highest probability
        class_name = class_names[np.argmax(probs)]
        
        # Plot image with class name and prob in the title
        fig2.add_subplot(3,2,i+1)
        plt.imshow(cv2.cvtColor(image, cv2.COLOR_BGR2RGB))
        plt.title("Class: " + class_name + ", probability: %.4f" %probs[0,np.argmax(probs)])
        plt.axis('off')