卷积神经网络中的参数数目计算

卷积神经网络中卷积层的参数设置与卷积核的大小还有数目有关。具体数目为：卷积核size×卷积核size×均积核的channel数目×卷积核的数目×bias(卷积核数目)；
全连接层的参数数目为：imput的matrix特征数目*全连接层Dense数值

import tensorflow as tf
from tensorflow.keras import layers, optimizers, datasets, Sequential
import os

os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'
tf.random.set_seed(2345)

conv_layers = [
    layers.Conv2D(64, kernel_size=[3, 3], padding='same', activation=tf.nn.relu),
    # 本层参数：卷积核size×卷积核size×均积核的channel数目×卷积核的数目×bias(卷积核数目)对应 3*3*3*64+64=1792
    layers.Conv2D(64, kernel_size=[3, 3], padding='same', activation=tf.nn.relu),
    #本层参数数目：3*3*64*64+64=36928，chanel数目为上一个卷积层的输出结果，为64
    layers.MaxPool2D(pool_size=[2, 2], strides=2, padding='same'),
    

    layers.Conv2D(128, kernel_size=[3, 3], padding="same", activation=tf.nn.relu),
    # 同理，本层参数为3*3*64*128+128=73856,64为上一个卷积层的输出为64
    layers.Conv2D(128, kernel_size=[3, 3], padding='same', activation=tf.nn.relu),
    layers.MaxPool2D(pool_size=[2, 2], strides=2, padding='same'),

    layers.Conv2D(256, kernel_size=[3, 3], padding="same", activation=tf.nn.relu),
    layers.Conv2D(256, kernel_size=[3, 3], padding='same', activation=tf.nn.relu),
    layers.MaxPool2D(pool_size=[2, 2], strides=2, padding='same'),

    layers.Conv2D(512, kernel_size=[3, 3], padding="same", activation=tf.nn.relu),
    layers.Conv2D(512, kernel_size=[3, 3], padding='same', activation=tf.nn.relu),
    layers.MaxPool2D(pool_size=[2, 2], strides=2, padding='same'),

    layers.Conv2D(512, kernel_size=[3, 3], padding="same", activation=tf.nn.relu),
    layers.Conv2D(512, kernel_size=[3, 3], padding='same', activation=tf.nn.relu),
    layers.MaxPool2D(pool_size=[2, 2], strides=2, padding='same'),
]


def preprocss(x, y):
    x = tf.cast(x, dtype=tf.float32)/255.
    y = tf.cast(y, dtype=tf.int32)
    return x, y

(x, y), (x_test, y_test) = datasets.cifar100.load_data()
y = tf.squeeze(y, axis=1)
y_test = tf.squeeze(y_test, axis=1)
print(x.shape, y.shape, x_test.shape, y_test.shape)

train_db = tf.data.Dataset.from_tensor_slices((x, y))
train_db = train_db.shuffle(1000).map(preprocss).batch(64)

train_test = tf.data.Dataset.from_tensor_slices((x_test, y_test))
train_test = train_test.map(preprocss).batch(64)

sample = next(iter(train_db))
print('sample: ', sample[0].shape, sample[1].shape,
      tf.reduce_min(sample[0]), tf.reduce_max(sample[0]))




def main():
    conv_net = Sequential(conv_layers)
    # x = tf.random.normal([4, 32, 32, 3])
    # out = conv_net(x)
    # print(out.shape)

    fc_net = Sequential([
        layers.Dense(256, activation=tf.nn.relu),#参数数目：imput dim（512）*256+256=131328
        layers.Dense(128, activation=tf.nn.relu),
        layers.Dense(100)
    ])

    conv_net.build(input_shape=[None, 32, 32, 3])
    fc_net.build(input_shape=[None, 512])

    conv_net.summary()
    fc_net.summary()


if __name__ == '__main__':
    main()

结果：
卷积层参数数目：

全连接层参数：