Hi I have a question in the process of quantizing to 4 bits. I created a new backend of get_default_qat_qconfig and set quant_min and quant_max from 0 to 15 to proceed with learning, but an error stating ‘zero_point’ must be between ‘quant_min’ and ‘quant_max’ appeared.
What else do I need to do to resolve this error?
Please refer to the picture below
2 Likes
I am having similar issue. can someone help with this?
can you print the zero point, quant_min and quant_max to see what i sthe problem?