How to apply 16-bits quantization to all the convolution layers in the model

Guides
1

You can use:
‘quantization_param({conv*}, precision_mode=a16_w16)\n’,
in the alls file.