“深度学习”学习日记：Tensorflow实现VGG每一个卷积层的可视化

2023.8.19

深度学习的卷积对于初学者是非常抽象，当时在入门学习的时候直接劝退一大班人，还好我坚持了下来。可视化时用到的图片（我们学校的一角！！！）以下展示了一个卷积和一次Relu的变化

作者使用的GPU是RTX 3050ti 在这张图像上已经出现了Cuda out of memory了。防止其他 图片出现类似情况：附上这张cat.jpg可以完成实验

代码是Copy大神的，用tensorflow1写的，使用tensoflow2的伙伴们，记得添上：

import tensorflow.compat.v1 as tf

Code:

# coding: utf-8

# # 使用预训练的VGG网络

# In[1]:

import scipy.io
import numpy as np
import os
import scipy.misc
import matplotlib.pyplot as plt
import tensorflow as tf
import imageio

import tensorflow.compat.v1 as tf


# get_ipython().magic(u'matplotlib inline')
print("所有包载入完毕")

# In[2]:

# 下载预先训练好的vgg-19模型，为Matlab的.mat格式，之后会用scipy读取
# (注意此版本模型与此处http://www.vlfeat.org/matconvnet/pretrained/最新版本不同)
import os.path

if not os.path.isfile('./data/imagenet-vgg-verydeep-19.mat'):
    os.system(
        u'wget -O data/imagenet-vgg-verydeep-19.mat http://www.vlfeat.org/matconvnet/models/beta16/imagenet-vgg-verydeep-19.mat')
    # get_ipython().system(u'wget -O data/imagenet-vgg-verydeep-19.mat http://www.vlfeat.org/matconvnet/models/beta16/imagenet-vgg-verydeep-19.mat')


# # 定义网络

# In[3]:

def net(data_path, input_image):
    layers = (
        'conv1_1', 'relu1_1', 'conv1_2', 'relu1_2', 'pool1',
        'conv2_1', 'relu2_1', 'conv2_2', 'relu2_2', 'pool2',
        'conv3_1', 'relu3_1', 'conv3_2', 'relu3_2', 'conv3_3',
        'relu3_3', 'conv3_4', 'relu3_4', 'pool3',
        'conv4_1', 'relu4_1', 'conv4_2', 'relu4_2', 'conv4_3',
        'relu4_3', 'conv4_4', 'relu4_4', 'pool4',
        'conv5_1', 'relu5_1', 'conv5_2', 'relu5_2', 'conv5_3',
        'relu5_3', 'conv5_4', 'relu5_4'
    )
    data = scipy.io.loadmat(data_path)
    mean_pixel = [103.939, 116.779, 123.68]
    weights = data['layers'][0]
    net = {}
    current = input_image
    for i, name in enumerate(layers):
        kind = name[:4]
        if kind == 'conv':
            kernels, bias = weights[i][0][0][0][0]
            # matconvnet: weights are [width, height, in_channels, out_channels]
            # tensorflow: weights are [height, width, in_channels, out_channels]
            kernels = np.transpose(kernels, (1, 0, 2, 3))
            bias = bias.reshape(-1)
            current = _conv_layer(current, kernels, bias)
        elif kind == 'relu':
            current = tf.nn.relu(current)
        elif kind == 'pool':
            current = _pool_layer(current)
        net[name] = current
    assert len(net) == len(layers)
    return net, mean_pixel, layers


print("Network for VGG ready")


# # 定义模型

# In[4]:

def _conv_layer(input, weights, bias):
    conv = tf.nn.conv2d(input, tf.constant(weights), strides=(1, 1, 1, 1),
                        padding='SAME')
    return tf.nn.bias_add(conv, bias)


def _pool_layer(input):
    return tf.nn.max_pool(input, ksize=(1, 2, 2, 1), strides=(1, 2, 2, 1),
                          padding='SAME')


def preprocess(image, mean_pixel):
    return image - mean_pixel


def unprocess(image, mean_pixel):
    return image + mean_pixel


def imread(path):
    # return scipy.misc.imread(path).astype(np.float)
    return imageio.imread(path)


def imsave(path, img):
    img = np.clip(img, 0, 255).astype(np.uint8)
    scipy.misc.imsave(path, img)


print("Functions for VGG ready")

# # 运行

# In[5]:

cwd = os.getcwd()
VGG_PATH = cwd + "/data/imagenet-vgg-verydeep-19.mat"
IMG_PATH = cwd + "/images/cat.jpg"
input_image = imread(IMG_PATH)
shape = (1,) + input_image.shape  # (h, w, nch) =>  (1, h, w, nch)
with tf.Graph().as_default(), tf.Session() as sess:
    image = tf.placeholder('float', shape=shape)
    nets, mean_pixel, all_layers = net(VGG_PATH, image)
    input_image_pre = np.array([preprocess(input_image, mean_pixel)])
    layers = all_layers  # For all layers
    # layers = ('relu2_1', 'relu3_1', 'relu4_1')
    for i, layer in enumerate(layers):
        print("[%d/%d] %s" % (i + 1, len(layers), layer))
        features = nets[layer].eval(feed_dict={image: input_image_pre})

        print(" Type of 'features' is ", type(features))
        print(" Shape of 'features' is %s" % (features.shape,))
        # Plot response 
        if 1:
            plt.figure(i + 1, figsize=(10, 5))
            plt.matshow(features[0, :, :, 0], cmap=plt.cm.gray, fignum=i + 1)
            plt.title("" + layer)
            plt.colorbar()
            plt.show()