FitResult.cost returns half the expected value of the objective function

[luisfabib]

The FitResult.cost attribute of the fit result objects should be returning the value of the objective function:

cost = sum(residual**2)

which right now is taken directly from the output of the different least-squares solvers in Scipy. However, these return half of the value. For example the least_squares function returns (as per the documentation):

cost = 1/2*sum(residual**2)

This must be corrected and made consistent throughout DeerLab.