在Pytorch中计算自己模型的FLOPs方式
时间:2020-10-20 13:22:03|栏目:Python代码|点击: 次
https://github.com/Lyken17/pytorch-OpCounter
安装方法很简单:
pip install thop
基本用法:
from torchvision.models import resnet50from thop import profile model = resnet50() flops, params = profile(model, input_size=(1, 3, 224,224))
对自己的module进行特别的计算:
class YourModule(nn.Module): # your definition def count_your_model(model, x, y): # your rule hereflops, params = profile(model, input_size=(1, 3, 224,224), custom_ops={YourModule: count_your_model})