MACs 和 FLOPs 有什么区别？如何计算？

MACs 和 FLOPs 有什么区别？如何计算？ 一、MACs 和 FLOPs 有什么区别？

FLOPs is abbreviation of floating operations which includes mul / add / div … etc.

MACs stands for multiply–accumulate operation that performs a <- a + (b x c).¹²

二、如何计算？

安装thop

pip install thop

pytorch使用thop：³

from torchvision.models import resnet50
from thop import profile
model = resnet50()
input = torch.randn(1, 3, 224, 224)
macs, params = profile(model, inputs=(input, ))

三、MACs 和 FLOPs的关系

粗略的：^4,1,5

MACs = 2 * FLOPs

---

1. https://github.com/Lyken17/pytorch-OpCounter/tree/master/benchmark#macs-flops-what-is-the-difference ↩︎ ↩︎

2. https://github.com/Lyken17/pytorch-OpCounter/issues/46 ↩︎

3. https://github.com/Lyken17/pytorch-OpCounter ↩︎

4. https://zhuanlan.zhihu.com/p/364543528 ↩︎

5. https://github.com/sovrasov/flops-counter.pytorch/issues/16 ↩︎