missing square in python implementation?

[yfarjoun]

Hello,

Thanks for the paper and the implementation!! It really makes it much easier to put into use.

Having read the paper and the code here I'm wondering about what seems to be a missing "square" around the term tf.maximum(0.,(1-alpha_) - PICP_S) in the python implementation.

Am I missing something?

[yfarjoun]

to be clear, I meant in the python notebook, not the python code!

[ZhangTeng-Hust]

I agree with you， but when I change it, I find the PICP <0.95.