程序師世界是廣大編程愛好者互助、分享、學習的平台,程序師世界有你更精彩!
首頁
編程語言
C語言|JAVA編程
Python編程
網頁編程
ASP編程|PHP編程
JSP編程
數據庫知識
MYSQL數據庫|SqlServer數據庫
Oracle數據庫|DB2數據庫
您现在的位置: 程式師世界 >> 編程語言 >  >> 更多編程語言 >> Python

Deep learning with Python cat dog recognition model, calculation process

編輯:Python

The calculation formula of convolution network is :
N=(W-F+2P)/S+1
among
N: Output size
W: Input size
F: Convolution kernel size
P: The size of the filling value
S: Step size

conv2d_1

Input [150,150,3] Output [148,148,32] The convolution kernel size is [3,3,32]

148=[150-3+2P]/S+1 Solution S=1,P=0

Max_pool2d_1 The convolution kernel size is [2,2,32]

Input [148,148,32] Pooling layer output [74,74,32]

74=[148-2+2P]/S+1 Step size of solution S=2, fill P=0

conv2d_2

Input [74,74,32], Output [72,72,64] The convolution kernel size is [3,3,64]

72=[74-3+2p]/s+1    step S=1, fill P=0

Max_pool2d_2 The convolution kernel size is [2,2,64]

Output [72,72,64], Output [36,36,64]

36=[72-2+2p]/s+1 step S=2, fill P=0

conv2d_3 The convolution kernel size is [3,3,128]

Output [36,36,64], Output [34,34,128]

34=[36-3+2p]/s+1  S=1 P=0

Max_pool2d_3 The convolution kernel size is [2,2,128]

Output [34,34,128], Output [17,17,128]

17=[34-2+2p]/s+1  S=2  P=0

conv2d_4   The convolution kernel size is [3,3,128]

Output [17,17,128], Output [15,15,128]

15=[17-3+2p]/s+1  S=1 P=0

Max_pool2d_4 The convolution kernel size is [2,2,128]

Output [15,15,128], Output [7,7,128]

(15+2p-2)/s+1 Integer =7 therefore   S=2 P=0


  1. 上一篇文章:
  2. 下一篇文章:
Copyright © 程式師世界 All Rights Reserved