vgg 16模型的内存和参数量的计算

左手的ㄟ右手 2023-06-27 06:11 1阅读 0赞

[   cs231n][cs231n]上关于VGG-16模型的内存和参数的计算过程如下。

INPUT: [224x224x3]        memory:  224*224*3=150K   weights: 0
    CONV3-64: [224x224x64]  memory:  224*224*64=3.2M   weights: (3*3*3)*64 = 1,728
    CONV3-64: [224x224x64]  memory:  224*224*64=3.2M   weights: (3*3*64)*64 = 36,864
    POOL2: [112x112x64]  memory:  112*112*64=800K   weights: 0
    CONV3-128: [112x112x128]  memory:  112*112*128=1.6M   weights: (3*3*64)*128 = 73,728
    CONV3-128: [112x112x128]  memory:  112*112*128=1.6M   weights: (3*3*128)*128 = 147,456
    POOL2: [56x56x128]  memory:  56*56*128=400K   weights: 0
    CONV3-256: [56x56x256]  memory:  56*56*256=800K   weights: (3*3*128)*256 = 294,912
    CONV3-256: [56x56x256]  memory:  56*56*256=800K   weights: (3*3*256)*256 = 589,824
    CONV3-256: [56x56x256]  memory:  56*56*256=800K   weights: (3*3*256)*256 = 589,824
    POOL2: [28x28x256]  memory:  28*28*256=200K   weights: 0
    CONV3-512: [28x28x512]  memory:  28*28*512=400K   weights: (3*3*256)*512 = 1,179,648
    CONV3-512: [28x28x512]  memory:  28*28*512=400K   weights: (3*3*512)*512 = 2,359,296
    CONV3-512: [28x28x512]  memory:  28*28*512=400K   weights: (3*3*512)*512 = 2,359,296
    POOL2: [14x14x512]  memory:  14*14*512=100K   weights: 0
    CONV3-512: [14x14x512]  memory:  14*14*512=100K   weights: (3*3*512)*512 = 2,359,296
    CONV3-512: [14x14x512]  memory:  14*14*512=100K   weights: (3*3*512)*512 = 2,359,296
    CONV3-512: [14x14x512]  memory:  14*14*512=100K   weights: (3*3*512)*512 = 2,359,296
    POOL2: [7x7x512]  memory:  7*7*512=25K  weights: 0
    FC: [1x1x4096]  memory:  4096  weights: 7*7*512*4096 = 102,760,448
    FC: [1x1x4096]  memory:  4096  weights: 4096*4096 = 16,777,216
    FC: [1x1x1000]  memory:  1000 weights: 4096*1000 = 4,096,000
    
    TOTAL memory: 24M * 4 bytes ~= 93MB / image (only forward! ~*2 for bwd)
    TOTAL params: 138M parameters

如果动手算一下memory的和，如下。

0.15M+3.2M+3.2M+0.8M+1.6M+1.6M+0.4M+0.8M3+0.2M+0.4M3+0.1M*4+0.025M+0.004M+0.004M+0.001M = 15.184M

问题来了，计算得到的总内存为15.184\*4 bytes ~=60MB / image，而教程中给的为93MB。并且教程中的数据(24M)被许多博客和文章所引用，导致经常看到的就是24M这个数据。起始该答案早在[github上有针对教程中的疑问][github]，并给出了如上手动计算的结果。所以，15M的memory才是正确的答案。    另外上述针对参数(weight)的计算中并没有加入bias的数量。

延伸1：模型的组成

参数的数量约为138M(不包含bias)，此时占用的内存大小为：138M\*4 bytes ~=526M，该大小约等于保存后模型占用磁盘的大小，而实际利用ImageNet训练出来的VGG-16的模型大小超过552M。那么差的20多M内存哪去了，此处不局限于该模型的探讨，而是想更一般化的探讨模型中的成分。

由上可知，模型中主要是权重和偏移单元，如vgg-16[加上偏移向量后的参数计算结果][Link 1]如下。然后是优化器和其他特殊层中的参数，如LeakyReLU，BN，和Dropout等的参数，如keras中的[model][]。

conv3-64  x 2       : 38,720
    conv3-128 x 2       : 221,440
    conv3-256 x 3       : 1,475,328
    conv3-512 x 3       : 5,899,776
    conv3-512 x 3       : 7,079,424
    fc1                 : 102,764,544
    fc2                 : 16,781,312
    fc3                 : 4,097,000
    TOTAL               : 138,357,544

其中全连接的计算为：

fc1 (x): (512x7x7)x4,096 (weights) + 4,096 (biases)
     fc2    : 4,096x4,096     (weights) + 4,096 (biases)
     fc3    : 4,096x1,000     (weights) + 1,000 (biases)

延伸2：训练和测试时的内存组成

![20180502141203789][]

图片来自[这里][Link 2]

[cs231n]: http://cs231n.github.io/convolutional-networks/#comp
[github]: https://github.com/cs231n/cs231n.github.io/issues/88
[Link 1]: https://stackoverflow.com/questions/28232235/how-to-calculate-the-number-of-parameters-of-convolutional-neural-networks
[model]: https://keras.io/models/about-keras-models/#about-keras-models
[20180502141203789]: https://img-blog.csdn.net/20180502141203789
[Link 2]: http://imatge-upc.github.io/telecombcn-2016-dlcv/slides/D2L1-memory.pdf