卷积速查表
作者Lou Xiao创建时间2020-09-18 18:48:00更新时间2022-11-01 18:58:00
深度学习(deep learning)中，常常用到2D卷积，尺寸计算就是一个问题，本文列出常用的2D卷积参数。
事实上只有输入与输出尺寸等于1:1是通用的，其它的比例都和input size相关的，故只列出比例是1：1的卷积参数据。
输入等于输出
注：dilation=1
| kernel size | stride | padding |
|---|---|---|
| 1 | 1 | 0 |
| 3 | 1 | 1 |
| 5 | 1 | 2 |
| 7 | 1 | 3 |
| 9 | 1 | 4 |
| 11 | 1 | 5 |
| 13 | 1 | 6 |
| 15 | 1 | 7 |
| 17 | 1 | 8 |
| 19 | 1 | 9 |
| 21 | 1 | 10 |
| 23 | 1 | 11 |
| 25 | 1 | 12 |
| 27 | 1 | 13 |
| 29 | 1 | 14 |
| 31 | 1 | 15 |
| 33 | 1 | 16 |
| 35 | 1 | 17 |
| 37 | 1 | 18 |
| 39 | 1 | 19 |
| 41 | 1 | 20 |
| 43 | 1 | 21 |
| 45 | 1 | 22 |
| 47 | 1 | 23 |
| 49 | 1 | 24 |
| 51 | 1 | 25 |
| 53 | 1 | 26 |
| 55 | 1 | 27 |
| 57 | 1 | 28 |
| 59 | 1 | 29 |
| 61 | 1 | 30 |
| 63 | 1 | 31 |
| 65 | 1 | 32 |
| 67 | 1 | 33 |
| 69 | 1 | 34 |
| 71 | 1 | 35 |
| 73 | 1 | 36 |
| 75 | 1 | 37 |
| 77 | 1 | 38 |
| 79 | 1 | 39 |
| 81 | 1 | 40 |
| 83 | 1 | 41 |
| 85 | 1 | 42 |
| 87 | 1 | 43 |
| 89 | 1 | 44 |
| 91 | 1 | 45 |
| 93 | 1 | 46 |
| 95 | 1 | 47 |
| 97 | 1 | 48 |
| 99 | 1 | 49 |
| 101 | 1 | 50 |
| 103 | 1 | 51 |
| 105 | 1 | 52 |
| 107 | 1 | 53 |
| 109 | 1 | 54 |
| 111 | 1 | 55 |
| 113 | 1 | 56 |
| 115 | 1 | 57 |
| 117 | 1 | 58 |
| 119 | 1 | 59 |
| 121 | 1 | 60 |
| 123 | 1 | 61 |
| 125 | 1 | 62 |
| 127 | 1 | 63 |
| 129 | 1 | 64 |
| 131 | 1 | 65 |
| 133 | 1 | 66 |
| 135 | 1 | 67 |
| 137 | 1 | 68 |
| 139 | 1 | 69 |
| 141 | 1 | 70 |
| 143 | 1 | 71 |
| 145 | 1 | 72 |
| 147 | 1 | 73 |
| 149 | 1 | 74 |
| 151 | 1 | 75 |
| 153 | 1 | 76 |
| 155 | 1 | 77 |
| 157 | 1 | 78 |
| 159 | 1 | 79 |
| 161 | 1 | 80 |
| 163 | 1 | 81 |
| 165 | 1 | 82 |
| 167 | 1 | 83 |
| 169 | 1 | 84 |
| 171 | 1 | 85 |
| 173 | 1 | 86 |
| 175 | 1 | 87 |
| 177 | 1 | 88 |
| 179 | 1 | 89 |
| 181 | 1 | 90 |
| 183 | 1 | 91 |
| 185 | 1 | 92 |
| 187 | 1 | 93 |
| 189 | 1 | 94 |
| 191 | 1 | 95 |
| 193 | 1 | 96 |
| 195 | 1 | 97 |
| 197 | 1 | 98 |
| 199 | 1 | 99 |
参考
文章目录