/
compute_flops.py
131 lines (103 loc) · 4.45 KB
/
compute_flops.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
# Code from https://github.com/simochen/model-tools.
import numpy as np
import os
import random
import shutil
import time
import warnings
import torch
import torchvision
import torch.nn as nn
from torch.autograd import Variable
import torch.nn.parallel
import torch.backends.cudnn as cudnn
import torch.distributed as dist
import torch.optim
import torch.utils.data
import torch.utils.data.distributed
import torchvision.transforms as transforms
import torchvision.datasets as datasets
import torchvision.models as models
def print_model_param_nums(model=None):
if model == None:
model = torchvision.models.alexnet()
total = sum([param.nelement() if param.requires_grad else 0 for param in model.parameters()])
print(' + Number of params: %.2fM' % (total / 1e6))
def count_model_param_flops(model=None, input_res=224, multiply_adds=True):
prods = {}
def save_hook(name):
def hook_per(self, input, output):
prods[name] = np.prod(input[0].shape)
return hook_per
list_1=[]
def simple_hook(self, input, output):
list_1.append(np.prod(input[0].shape))
list_2={}
def simple_hook2(self, input, output):
list_2['names'] = np.prod(input[0].shape)
list_conv=[]
def conv_hook(self, input, output):
batch_size, input_channels, input_height, input_width = input[0].size()
output_channels, output_height, output_width = output[0].size()
kernel_ops = self.kernel_size[0] * self.kernel_size[1] * (self.in_channels / self.groups)
bias_ops = 1 if self.bias is not None else 0
params = output_channels * (kernel_ops + bias_ops)
num_weight_params = (self.weight.data != 0).float().sum()
assert self.weight.numel() == kernel_ops * output_channels, "Not match"
flops = (num_weight_params * (2 if multiply_adds else 1) + bias_ops * output_channels) * output_height * output_width * batch_size
list_conv.append(flops)
list_linear=[]
def linear_hook(self, input, output):
batch_size = input[0].size(0) if input[0].dim() == 2 else 1
weight_ops = self.weight.nelement() * (2 if multiply_adds else 1)
bias_ops = self.bias.nelement()
flops = batch_size * (weight_ops + bias_ops)
list_linear.append(flops)
list_bn=[]
def bn_hook(self, input, output):
list_bn.append(input[0].nelement() * 2)
list_relu=[]
def relu_hook(self, input, output):
list_relu.append(input[0].nelement())
list_pooling=[]
def pooling_hook(self, input, output):
batch_size, input_channels, input_height, input_width = input[0].size()
output_channels, output_height, output_width = output[0].size()
kernel_ops = self.kernel_size * self.kernel_size
bias_ops = 0
params = 0
flops = (kernel_ops + bias_ops) * output_channels * output_height * output_width * batch_size
list_pooling.append(flops)
list_upsample=[]
# For bilinear upsample
def upsample_hook(self, input, output):
batch_size, input_channels, input_height, input_width = input[0].size()
output_channels, output_height, output_width = output[0].size()
flops = output_height * output_width * output_channels * batch_size * 12
list_upsample.append(flops)
def foo(net):
childrens = list(net.children())
if not childrens:
if isinstance(net, torch.nn.Conv2d):
net.register_forward_hook(conv_hook)
if isinstance(net, torch.nn.Linear):
net.register_forward_hook(linear_hook)
if isinstance(net, torch.nn.BatchNorm2d):
net.register_forward_hook(bn_hook)
if isinstance(net, torch.nn.ReLU):
net.register_forward_hook(relu_hook)
if isinstance(net, torch.nn.MaxPool2d) or isinstance(net, torch.nn.AvgPool2d):
net.register_forward_hook(pooling_hook)
if isinstance(net, torch.nn.Upsample):
net.register_forward_hook(upsample_hook)
return
for c in childrens:
foo(c)
if model == None:
model = torchvision.models.alexnet()
foo(model)
input = Variable(torch.rand(3,input_res,input_res).unsqueeze(0), requires_grad = True)
out = model(input)
total_flops = (sum(list_conv) + sum(list_linear) + sum(list_bn) + sum(list_relu) + sum(list_pooling) + sum(list_upsample))
print(' + Number of FLOPs: %.2fG' % (total_flops / 1e9))
return total_flops