Guidance for QAT

[Ba1tu3han]

Hello everyone,

I am trying to quantize a pre-trained model (mobilenet_v3_small) by using QAT of Brevitas to binary.

I could not find to correct starting point. Could you lead me?

Can I convert all layers to quantized version by automatically or have to write code manually by hand?

Actually, I want a basic script that loads my model then I configure some fundamental parameters and datasets, and it gives me the quantized model. Is it possible? Because I have limited coding and linux knowledge.

Best Regards,

Batuhan