8bit quantization calibrator and fine-tuning

[Hong-333]

Hi, I am new to intel caffe.

I am interested in 8bit quantization. I saw this "Introduction of Accuracy Calibration Tool for 8 Bit Inference"

I got the result using calibrator 'Failed to get accuracy, please check the parameters and rerun the scripts.' is that error?

and what is scale_in, scale_out,scale_param

Finally I want to know After 8-bit quantization can fine-tuning using quantized.prototxt to increase accuracy?

Thank you!