一行代码计算模型的参数量和FLOPs【亲测有效】

安装对应的库文件

pip install ptflops

代码（get_model_complexity_info)

import torchvision.models as models
from ptflops import get_model_complexity_info
net = models.vgg16() #可以为自己搭建的模型
flops, params = get_model_complexity_info(model, (3,512,512), as_strings=True, print_per_layer_stat=True)  #(3,512,512)输入图片的尺寸
print("Flops: {}".format(flops))
print("Params: " + params)