What does it mean if Gauss-Newton converges in one iteration with a particular choice of initialisation?

128 Views Asked by At

I have a nonlinear least squares optimisation problem. With a particular choice of initialisation, it appears that the Gauss-Newton method always converges in one iteration for my problem. With a random initialisation, Gauss-Newton appears to still converge to the same solution but takes more iterations.

What does this say about my residual function?