If my training values are a 0 or a 1, what loss function should I use?

Question

So, if my training values look like this

X_train

[[[0.03699663]]

 [[0.03699663]]

 [[0.03722753]]

 ...

 [[0.71153342]]

 [[0.71062594]]

 [[0.71067825]]]

y_t