could you try this solution below?
It means adding quantization_param([conv42, conv53, conv63], force_range_out=[0.0, 1.0]) to this file:
quantization_param([conv42, conv53, conv63], force_range_out=[0.0, 1.0])