直接print就可以打印网络结构
model = resnet50(pretrained=True)
print(model)
但是这样有一点乱，如果结构复杂的话
torchinfo
安装这个库
pip install torchinfo
import torchinfo
model = resnet50(pretrained=True)
# 设置input_size 因为会为你计算参数量
torchinfo.summary(model=model,input_size=(16, 3, 224, 224))
打印结果如下：
==========================================================================================
Layer (type:depth-idx) Output Shape Param #
==========================================================================================
ResNet [16, 1000] --
├─Conv2d: 1-1 [16, 64, 112, 112] 9,408
├─BatchNorm2d: 1-2 [16, 64, 112, 112] 128
├─ReLU: 1-3 [16, 64, 112, 112] --
├─MaxPool2d: 1-4 [16, 64, 56, 56] --
├─Sequential: 1-5 [16, 256, 56, 56] --
│ └─Bottleneck: 2-1 [16, 256, 56, 56] --
│ │ └─Conv2d: 3-1 [16, 64, 56, 56] 4,096
│ │ └─BatchNorm2d: 3-2 [16, 64, 56, 56] 128
│ │ └─ReLU: 3-3 [16, 64, 56, 56] --
│ │ └─Conv2d: 3-4 [16, 64, 56, 56] 36,864
│ │ └─BatchNorm2d: 3-5 [16, 64, 56, 56] 128
│ │ └─ReLU: 3-6 [16, 64, 56, 56] --
│ │ └─Conv2d: 3-7 [16, 256, 56, 56] 16,384
│ │ └─BatchNorm2d: 3-8 [16, 256, 56, 56] 512
│ │ └─Sequential: 3-9 [16, 256, 56, 56] 16,896
│ │ └─ReLU: 3-10 [16, 256, 56, 56] --
│ └─Bottleneck: 2-2 [16, 256, 56, 56] --
│ │ └─Conv2d: 3-11 [16, 64, 56, 56] 16,384
│ │ └─BatchNorm2d: 3-12 [16, 64, 56, 56] 128
│ │ └─ReLU: 3-13 [16, 64, 56, 56] --
│ │ └─Conv2d: 3-14 [16, 64, 56, 56] 36,864
│ │ └─BatchNorm2d: 3-15 [16, 64, 56, 56] 128
│ │ └─ReLU: 3-16 [16, 64, 56, 56] --
│ │ └─Conv2d: 3-17 [16, 256, 56, 56] 16,384
│ │ └─BatchNorm2d: 3-18 [16, 256, 56, 56] 512
│ │ └─ReLU: 3-19 [16, 256, 56, 56] --
│ └─Bottleneck: 2-3 [16, 256, 56, 56] --
│ │ └─Conv2d: 3-20 [16, 64, 56, 56] 16,384
│ │ └─BatchNorm2d: 3-21 [16, 64, 56, 56] 128
│ │ └─ReLU: 3-22 [16, 64, 56, 56] --
│ │ └─Conv2d: 3-23 [16, 64, 56, 56] 36,864
│ │ └─BatchNorm2d: 3-24 [16, 64, 56, 56] 128
│ │ └─ReLU: 3-25 [16, 64, 56, 56] --
│ │ └─Conv2d: 3-26 [16, 256, 56, 56] 16,384
│ │ └─BatchNorm2d: 3-27 [16, 256, 56, 56] 512
│ │ └─ReLU: 3-28 [16, 256, 56, 56] --
├─Sequential: 1-6 [16, 512, 28, 28] --
│ └─Bottleneck: 2-4 [16, 512, 28, 28] --
│ │ └─Conv2d: 3-29 [16, 128, 56, 56] 32,768
│ │ └─BatchNorm2d: 3-30 [16, 128, 56, 56] 256
│ │ └─ReLU: 3-31 [16, 128, 56, 56] --
│ │ └─Conv2d: 3-32 [16, 128, 28, 28] 147,456
│ │ └─BatchNorm2d: 3-33 [16, 128, 28, 28] 256
│ │ └─ReLU: 3-34 [16, 128, 28, 28] --
│ │ └─Conv2d: 3-35 [16, 512, 28, 28] 65,536
│ │ └─BatchNorm2d: 3-36 [16, 512, 28, 28] 1,024
│ │ └─Sequential: 3-37 [16, 512, 28, 28] 132,096
│ │ └─ReLU: 3-38 [16, 512, 28, 28] --
│ └─Bottleneck: 2-5 [16, 512, 28, 28] --
│ │ └─Conv2d: 3-39 [16, 128, 28, 28] 65,536
│ │ └─BatchNorm2d: 3-40 [16, 128, 28, 28] 256
│ │ └─ReLU: 3-41 [16, 128, 28, 28] --
│ │ └─Conv2d: 3-42 [16, 128, 28, 28] 147,456
│ │ └─BatchNorm2d: 3-43 [16, 128, 28, 28] 256
│ │ └─ReLU: 3-44 [16, 128, 28, 28] --
│ │ └─Conv2d: 3-45 [16, 512, 28, 28] 65,536
│ │ └─BatchNorm2d: 3-46 [16, 512, 28, 28] 1,024
│ │ └─ReLU: 3-47 [16, 512, 28, 28] --
│ └─Bottleneck: 2-6 [16, 512, 28, 28] --
│ │ └─Conv2d: 3-48 [16, 128, 28, 28] 65,536
│ │ └─BatchNorm2d: 3-49 [16, 128, 28, 28] 256
│ │ └─ReLU: 3-50 [16, 128, 28, 28] --
│ │ └─Conv2d: 3-51 [16, 128, 28, 28] 147,456
│ │ └─BatchNorm2d: 3-52 [16, 128, 28, 28] 256
│ │ └─ReLU: 3-53 [16, 128, 28, 28] --
│ │ └─Conv2d: 3-54 [16, 512, 28, 28] 65,536
│ │ └─BatchNorm2d: 3-55 [16, 512, 28, 28] 1,024
│ │ └─ReLU: 3-56 [16, 512, 28, 28] --
│ └─Bottleneck: 2-7 [16, 512, 28, 28] --
│ │ └─Conv2d: 3-57 [16, 128, 28, 28] 65,536
│ │ └─BatchNorm2d: 3-58 [16, 128, 28, 28] 256
│ │ └─ReLU: 3-59 [16, 128, 28, 28] --
│ │ └─Conv2d: 3-60 [16, 128, 28, 28] 147,456
│ │ └─BatchNorm2d: 3-61 [16, 128, 28, 28] 256
│ │ └─ReLU: 3-62 [16, 128, 28, 28] --
│ │ └─Conv2d: 3-63 [16, 512, 28, 28] 65,536
│ │ └─BatchNorm2d: 3-64 [16, 512, 28, 28] 1,024
│ │ └─ReLU: 3-65 [16, 512, 28, 28] --
├─Sequential: 1-7 [16, 1024, 14, 14] --
│ └─Bottleneck: 2-8 [16, 1024, 14, 14] --
│ │ └─Conv2d: 3-66 [16, 256, 28, 28] 131,072
│ │ └─BatchNorm2d: 3-67 [16, 256, 28, 28] 512
│ │ └─ReLU: 3-68 [16, 256, 28, 28] --
│ │ └─Conv2d: 3-69 [16, 256, 14, 14] 589,824
│ │ └─BatchNorm2d: 3-70 [16, 256, 14, 14] 512
│ │ └─ReLU: 3-71 [16, 256, 14, 14] --
│ │ └─Conv2d: 3-72 [16, 1024, 14, 14] 262,144
│ │ └─BatchNorm2d: 3-73 [16, 1024, 14, 14] 2,048
│ │ └─Sequential: 3-74 [16, 1024, 14, 14] 526,336
│ │ └─ReLU: 3-75 [16, 1024, 14, 14] --
│ └─Bottleneck: 2-9 [16, 1024, 14, 14] --
│ │ └─Conv2d: 3-76 [16, 256, 14, 14] 262,144
│ │ └─BatchNorm2d: 3-77 [16, 256, 14, 14] 512
│ │ └─ReLU: 3-78 [16, 256, 14, 14] --
│ │ └─Conv2d: 3-79 [16, 256, 14, 14] 589,824
│ │ └─BatchNorm2d: 3-80 [16, 256, 14, 14] 512
│ │ └─ReLU: 3-81 [16, 256, 14, 14] --
│ │ └─Conv2d: 3-82 [16, 1024, 14, 14] 262,144
│ │ └─BatchNorm2d: 3-83 [16, 1024, 14, 14] 2,048
│ │ └─ReLU: 3-84 [16, 1024, 14, 14] --
│ └─Bottleneck: 2-10 [16, 1024, 14, 14] --
│ │ └─Conv2d: 3-85 [16, 256, 14, 14] 262,144
│ │ └─BatchNorm2d: 3-86 [16, 256, 14, 14] 512
│ │ └─ReLU: 3-87 [16, 256, 14, 14] --
│ │ └─Conv2d: 3-88 [16, 256, 14, 14] 589,824
│ │ └─BatchNorm2d: 3-89 [16, 256, 14, 14] 512
│ │ └─ReLU: 3-90 [16, 256, 14, 14] --
│ │ └─Conv2d: 3-91 [16, 1024, 14, 14] 262,144
│ │ └─BatchNorm2d: 3-92 [16, 1024, 14, 14] 2,048
│ │ └─ReLU: 3-93 [16, 1024, 14, 14] --
│ └─Bottleneck: 2-11 [16, 1024, 14, 14] --
│ │ └─Conv2d: 3-94 [16, 256, 14, 14] 262,144
│ │ └─BatchNorm2d: 3-95 [16, 256, 14, 14] 512
│ │ └─ReLU: 3-96 [16, 256, 14, 14] --
│ │ └─Conv2d: 3-97 [16, 256, 14, 14] 589,824
│ │ └─BatchNorm2d: 3-98 [16, 256, 14, 14] 512
│ │ └─ReLU: 3-99 [16, 256, 14, 14] --
│ │ └─Conv2d: 3-100 [16, 1024, 14, 14] 262,144
│ │ └─BatchNorm2d: 3-101 [16, 1024, 14, 14] 2,048
│ │ └─ReLU: 3-102 [16, 1024, 14, 14] --
│ └─Bottleneck: 2-12 [16, 1024, 14, 14] --
│ │ └─Conv2d: 3-103 [16, 256, 14, 14] 262,144
│ │ └─BatchNorm2d: 3-104 [16, 256, 14, 14] 512
│ │ └─ReLU: 3-105 [16, 256, 14, 14] --
│ │ └─Conv2d: 3-106 [16, 256, 14, 14] 589,824
│ │ └─BatchNorm2d: 3-107 [16, 256, 14, 14] 512
│ │ └─ReLU: 3-108 [16, 256, 14, 14] --
│ │ └─Conv2d: 3-109 [16, 1024, 14, 14] 262,144
│ │ └─BatchNorm2d: 3-110 [16, 1024, 14, 14] 2,048
│ │ └─ReLU: 3-111 [16, 1024, 14, 14] --
│ └─Bottleneck: 2-13 [16, 1024, 14, 14] --
│ │ └─Conv2d: 3-112 [16, 256, 14, 14] 262,144
│ │ └─BatchNorm2d: 3-113 [16, 256, 14, 14] 512
│ │ └─ReLU: 3-114 [16, 256, 14, 14] --
│ │ └─Conv2d: 3-115 [16, 256, 14, 14] 589,824
│ │ └─BatchNorm2d: 3-116 [16, 256, 14, 14] 512
│ │ └─ReLU: 3-117 [16, 256, 14, 14] --
│ │ └─Conv2d: 3-118 [16, 1024, 14, 14] 262,144
│ │ └─BatchNorm2d: 3-119 [16, 1024, 14, 14] 2,048
│ │ └─ReLU: 3-120 [16, 1024, 14, 14] --
├─Sequential: 1-8 [16, 2048, 7, 7] --
│ └─Bottleneck: 2-14 [16, 2048, 7, 7] --
│ │ └─Conv2d: 3-121 [16, 512, 14, 14] 524,288
│ │ └─BatchNorm2d: 3-122 [16, 512, 14, 14] 1,024
│ │ └─ReLU: 3-123 [16, 512, 14, 14] --
│ │ └─Conv2d: 3-124 [16, 512, 7, 7] 2,359,296
│ │ └─BatchNorm2d: 3-125 [16, 512, 7, 7] 1,024
│ │ └─ReLU: 3-126 [16, 512, 7, 7] --
│ │ └─Conv2d: 3-127 [16, 2048, 7, 7] 1,048,576
│ │ └─BatchNorm2d: 3-128 [16, 2048, 7, 7] 4,096
│ │ └─Sequential: 3-129 [16, 2048, 7, 7] 2,101,248
│ │ └─ReLU: 3-130 [16, 2048, 7, 7] --
│ └─Bottleneck: 2-15 [16, 2048, 7, 7] --
│ │ └─Conv2d: 3-131 [16, 512, 7, 7] 1,048,576
│ │ └─BatchNorm2d: 3-132 [16, 512, 7, 7] 1,024
│ │ └─ReLU: 3-133 [16, 512, 7, 7] --
│ │ └─Conv2d: 3-134 [16, 512, 7, 7] 2,359,296
│ │ └─BatchNorm2d: 3-135 [16, 512, 7, 7] 1,024
│ │ └─ReLU: 3-136 [16, 512, 7, 7] --
│ │ └─Conv2d: 3-137 [16, 2048, 7, 7] 1,048,576
│ │ └─BatchNorm2d: 3-138 [16, 2048, 7, 7] 4,096
│ │ └─ReLU: 3-139 [16, 2048, 7, 7] --
│ └─Bottleneck: 2-16 [16, 2048, 7, 7] --
│ │ └─Conv2d: 3-140 [16, 512, 7, 7] 1,048,576
│ │ └─BatchNorm2d: 3-141 [16, 512, 7, 7] 1,024
│ │ └─ReLU: 3-142 [16, 512, 7, 7] --
│ │ └─Conv2d: 3-143 [16, 512, 7, 7] 2,359,296
│ │ └─BatchNorm2d: 3-144 [16, 512, 7, 7] 1,024
│ │ └─ReLU: 3-145 [16, 512, 7, 7] --
│ │ └─Conv2d: 3-146 [16, 2048, 7, 7] 1,048,576
│ │ └─BatchNorm2d: 3-147 [16, 2048, 7, 7] 4,096
│ │ └─ReLU: 3-148 [16, 2048, 7, 7] --
├─AdaptiveAvgPool2d: 1-9 [16, 2048, 1, 1] --
├─Linear: 1-10 [16, 1000] 2,049,000
==========================================================================================
Total params: 25,557,032
Trainable params: 25,557,032
Non-trainable params: 0
Total mult-adds (G): 65.43
==========================================================================================
Input size (MB): 9.63
Forward/backward pass size (MB): 2845.31
Params size (MB): 102.23
Estimated Total Size (MB): 2957.17
==========================================================================================