网络模型的参数量和FLOPs的计算 Pytorch(网络模型参数方法)

网络模型的参数量和FLOPs的计算 Pytorch

目录

1、torchstat 

2、thop

3、fvcore 

4、flops_counter

5、自定义统计函数

---

推荐整理分享网络模型的参数量和FLOPs的计算 Pytorch(网络模型参数方法)，希望有所帮助，仅作参考，欢迎阅读内容。

文章相关热门搜索词:网络模型的参数量是一层不变的吗,网络模型的参数设置,网络模型的参数是什么,网络模型参数量如何计算,网络模型的参数量,网络模型的参数是什么,网络模型的参数量是一层不变的吗,网络模型的参数有哪些,内容如对您有帮助，希望把文章链接给更多的朋友！

FLOPS和FLOPs的区别：

FLOPS：注意全大写，是floating point operations per second的缩写，意指每秒浮点运算次数，理解为计算速度。是一个衡量硬件性能的指标。FLOPs：注意s小写，是floating point operations的缩写（s表复数），意指浮点运算数，理解为计算量。可以用来衡量算法/模型的复杂度。

在介绍torchstat包和thop包之前，先总结一下：

torchstat包可以统计卷积神经网络和全连接神经网络的参数和计算量。thop包可以统计统计卷积神经网络、全连接神经网络以及循环神经网络的参数和计算量，程序示例等详见下文。1、torchstat pip install torchstat -i https://pypi.tuna.tsinghua.edu.cn/simple

在实际操作中，我们可以调用torchstat包，帮助我们统计模型的parameters和FLOPs。如果不修改这个包里面的一些代码，那么这个包只适用于输入为3通道的图像的模型。

import torchimport torch.nn as nnfrom torchstat import statclass Simple(nn.Module): def __init__(self): super().__init__() self.conv1 = nn.Conv2d(3, 16, 3, 1, padding=1, bias=False) self.conv2 = nn.Conv2d(16, 32, 3, 1, padding=1, bias=False) def forward(self, x): x = self.conv1(x) x = self.conv2(x) return xmodel = Simple()stat(model, (3, 244, 244)) # 统计模型的参数量和FLOPs，（3,244,244）是输入图像的size

 如果把torchstat包中的一行程序进行一点点改动，那么这个包可以用来统计全连接神经网络的参数量和计算量。当然手动计算全连接神经网络的参数量和计算量也很快 =_= 。进入torchstat源代码之后，如下图所示，注释掉圈红的地方，就可以用torchstat包统计全连接神经网络的参数量和计算量了。

2、thoppip install thop -i https://pypi.tuna.tsinghua.edu.cn/simpleimport torchimport torch.nn as nnfrom thop import profileclass Simple(nn.Module): def __init__(self): super().__init__() self.fc1 = nn.Linear(10, 10) def forward(self, x): x = self.fc1(x) return xnet = Simple()input = torch.randn(1, 10) # batchsize=1, 输入向量长度为10macs, params = profile(net, inputs=(input, ))print(' FLOPs: ', macs*2) # 一般来讲，FLOPs是macs的两倍print('params: ', params)3、fvcore pip install fvcore -i https://pypi.tuna.tsinghua.edu.cn/simple

用它比较好

import torchfrom torchvision.models import resnet50from fvcore.nn import FlopCountAnalysis, parameter_count_table# 创建resnet50网络model = resnet50(num_classes=1000)# 创建输入网络的tensortensor = (torch.rand(1, 3, 224, 224),)# 分析FLOPsflops = FlopCountAnalysis(model, tensor)print("FLOPs: ", flops.total())# 分析parametersprint(parameter_count_table(model))

 终端输出结果如下，FLOPs为4089184256，模型参数数量约为25.6M（这里的参数数量和我自己计算的有些出入，主要是在BN模块中，这里只计算了beta和gamma两个训练参数，没有统计moving_mean和moving_var两个参数），具体可以看下我在官方提的issue。 通过终端打印的信息我们可以发现在计算FLOPs时并没有包含BN层，池化层还有普通的add操作（我发现计算FLOPs时并没有统一的规定，在github上看的计算FLOPs项目基本每个都不同，但计算出来的结果大同小异）。

注意：在使用fvcore模块计算模型的flops时，遇到了问题，记录一下解决方案。首先是在jit_analysis.py的589行出错。经过调试发现，op_counts.values()的类型是int32，但是计算要求的类型只能是int、float、np.float64和np.int64，因此需要手动进行强制转换。修改如下：

4、flops_counterpip install ptflops -i https://pypi.tuna.tsinghua.edu.cn/simple

用它也很好，结果和fvcore一样

from ptflops import get_model_complexity_infomacs, params = get_model_complexity_info(model, (112, 9, 9), as_strings=True, print_per_layer_stat=True, verbose=True)print('{:<30} {:<8}'.format('Computational complexity: ', macs))print('{:<30} {:<8}'.format('Number of parameters: ', params))

5、自定义统计函数import torchimport numpy as npdef calc_flops(model, input): 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) * ( 2 if multiply_adds else 1) bias_ops = 1 if self.bias is not None else 0 params = output_channels * (kernel_ops + bias_ops) flops = batch_size * params * output_height * output_width list_conv.append(flops) def linear_hook(self, input, output): batch_size = input[0].size(0) if input[0].dim() == 2 else 1 num_steps = input[0].size(0) weight_ops = self.weight.nelement() * (2 if multiply_adds else 1) bias_ops = self.bias.nelement() if self.bias is not None else 0 flops = batch_size * (weight_ops + bias_ops) flops *= num_steps list_linear.append(flops) def fsmn_hook(self, input, output): batch_size = input[0].size(0) if input[0].dim() == 2 else 1 weight_ops = self.filter.nelement() * (2 if multiply_adds else 1) num_steps = input[0].size(0) flops = num_steps * weight_ops flops *= batch_size list_fsmn.append(flops) def gru_cell(input_size, hidden_size, bias=True): total_ops = 0 # r = \sigma(W_{ir} x + b_{ir} + W_{hr} h + b_{hr}) \\ # z = \sigma(W_{iz} x + b_{iz} + W_{hz} h + b_{hz}) \\ state_ops = (hidden_size + input_size) * hidden_size + hidden_size if bias: state_ops += hidden_size * 2 total_ops += state_ops * 2 # n = \tanh(W_{in} x + b_{in} + r * (W_{hn} h + b_{hn})) \\ total_ops += (hidden_size + input_size) * hidden_size + hidden_size if bias: total_ops += hidden_size * 2 # r hadamard : r * (~) total_ops += hidden_size # h' = (1 - z) * n + z * h # hadamard hadamard add total_ops += hidden_size * 3 return total_ops def gru_hook(self, input, output): batch_size = input[0].size(0) if input[0].dim() == 2 else 1 if self.batch_first: batch_size = input[0].size(0) num_steps = input[0].size(1) else: batch_size = input[0].size(1) num_steps = input[0].size(0) total_ops = 0 bias = self.bias input_size = self.input_size hidden_size = self.hidden_size num_layers = self.num_layers total_ops = 0 total_ops += gru_cell(input_size, hidden_size, bias) for i in range(num_layers - 1): total_ops += gru_cell(hidden_size, hidden_size, bias) total_ops *= batch_size total_ops *= num_steps list_lstm.append(total_ops) def lstm_cell(input_size, hidden_size, bias): total_ops = 0 state_ops = (input_size + hidden_size) * hidden_size + hidden_size if bias: state_ops += hidden_size * 2 total_ops += state_ops * 4 total_ops += hidden_size * 3 total_ops += hidden_size return total_ops def lstm_hook(self, input, output): batch_size = input[0].size(0) if input[0].dim() == 2 else 1 if self.batch_first: batch_size = input[0].size(0) num_steps = input[0].size(1) else: batch_size = input[0].size(1) num_steps = input[0].size(0) total_ops = 0 bias = self.bias input_size = self.input_size hidden_size = self.hidden_size num_layers = self.num_layers total_ops = 0 total_ops += lstm_cell(input_size, hidden_size, bias) for i in range(num_layers - 1): total_ops += lstm_cell(hidden_size, hidden_size, bias) total_ops *= batch_size total_ops *= num_steps list_lstm.append(total_ops) def bn_hook(self, input, output): list_bn.append(input[0].nelement()) def relu_hook(self, input, output): list_relu.append(input[0].nelement()) 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 = output_channels * (kernel_ops + bias_ops) flops = batch_size * params * output_height * output_width list_pooling.append(flops) def foo(net): childrens = list(net.children()) if not childrens: print(net) if isinstance(net, torch.nn.Conv2d) or isinstance(net, torch.nn.ConvTranspose2d): net.register_forward_hook(conv_hook) # print('conv_hook_ready') if isinstance(net, torch.nn.Linear): net.register_forward_hook(linear_hook) # print('linear_hook_ready') if isinstance(net, torch.nn.BatchNorm2d): net.register_forward_hook(bn_hook) # print('batch_norm_hook_ready') if isinstance(net, torch.nn.ReLU) or isinstance(net, torch.nn.PReLU): net.register_forward_hook(relu_hook) # print('relu_hook_ready') if isinstance(net, torch.nn.MaxPool2d) or isinstance(net, torch.nn.AvgPool2d): net.register_forward_hook(pooling_hook) # print('pooling_hook_ready') if isinstance(net, torch.nn.LSTM): net.register_forward_hook(lstm_hook) # print('lstm_hook_ready') if isinstance(net, torch.nn.GRU): net.register_forward_hook(gru_hook) # if isinstance(net, FSMNZQ): # net.register_forward_hook(fsmn_hook) # print('fsmn_hook_ready') return for c in childrens: foo(c) multiply_adds = False list_conv, list_bn, list_relu, list_linear, list_pooling, list_lstm, list_fsmn = [], [], [], [], [], [], [] foo(model) _ = model(input) total_flops = (sum(list_conv) + sum(list_linear) + sum(list_bn) + sum(list_relu) + sum(list_pooling) + sum( list_lstm) + sum(list_fsmn)) fsmn_flops = (sum(list_fsmn) + sum(list_linear)) lstm_flops = sum(list_lstm) model_parameters = filter(lambda p: p.requires_grad, model.parameters()) params = sum([np.prod(p.size()) for p in model_parameters]) print('The network has {} params.'.format(params)) print(total_flops, fsmn_flops, lstm_flops) print(' + Number of FLOPs: %.2f M' % (total_flops / 1000 ** 2)) return total_flopsif __name__ == '__main__': from torchvision.models import resnet18 model = resnet18(num_classes=1000) imput_size = torch.rand((1,3,224,224)) calc_flops(model, imput_size)