Open songqi-github opened 4 years ago
I meet the same problesms。
ResNet-50 has 4.12 GFLOPs (i.e., 4.12 billion floating-point operations). See: https://mmclassification.readthedocs.io/en/latest/model_zoo.html.
I get this exact value when I take resnet50
from torchvision.models
and calculate its FLOPs using this library.
When computing the GFLOPs of ResNet50, the answer is around 100GFLOPs. However, the result in other paper is around 500. Is there something wrong with your computing method?