I tried to use inference engine with I8 model precision on CPU, but I8 model could not be loaded.
This model could be loaded with FP32 model precision, but I8 could not.
Error message is "Supported primitive descriptors list is empty for node: LinearFunction_1"
Thank you for help.
Hi Cary P.
Thank you for replying.
I am using private model which is generated by client company.
I got Int8 IR model by command
$python3 calibration_tool.py -sm -m <FP32 IR model path>.xml --output-dir <output directory path>.
Are there some Int8 model which can not be loaded?
I'm sorry, with limited information, it's hard for me to help verify where the problem is. BUT there will be new calibration tool in R4, you can wait for the new tool and test again.