How to count the OPs of quantized models

I wonder how to quantize the OPs of quantized models, such as torchvision.models.quantization.resnet18.

I have tried torchstat, torchinfo… but the result seems to be wrong.
Thanks for your help!