Dynamic + Post-training static quantization at the same time

quantization
1

Hi,
I have a CNN+linear model that I want to quantize. To make the most of PyTorch’s quantization, is there any way to perform static quantization on the CNN and dynamic on linear?

Thank you.

2

if you are doing static quantization, why don’t do static quantization on both conv and linear?

If you really need to do this, you can first apply static quantization, and then dynamic quantization.