[*^*]模型参数和参数量计算 thop库
模型参数和参数量计算 thop库
THOP: PyTorch-OpCounter
How to install
pip install thop
(now continously intergrated on Github actions)
OR
pip install --upgrade git+https://github.com/Lyken17/pytorch-OpCounter.git
How to use
-
Basic usage
from torchvision.models import resnet50 from thop import profile model = resnet50() input = torch.randn(1, 3, 224, 224) macs, params = profile(model, inputs=(input, ))
-
Define the rule for 3rd party module.
class YourModule(nn.Module): # your definition def count_your_model(model, x, y): # your rule here input = torch.randn(1, 3, 224, 224) macs, params = profile(model, inputs=(input, ), custom_ops={YourModule: count_your_model})
-
Improve the output readability
Call
thop.clever_format
to give a better format of the output.from thop import clever_format macs, params = clever_format([macs, params], "%.3f")
Results of Recent Models
The implementation are adapted from torchvision
. Following results can be obtained using benchmark/evaluate_famous_models.py.
|
|