It seems that there lost a parentheses in the compute_hessian function. -- (p-n) / 2.*eps

[Frizy-up]

compute_hessian function last line :
hessian = [(p-n) / 2.*eps for p, n in zip(dalpha_pos, dalpha_neg)] should be changed to:
hessian = [(p-n) / (2.*eps) for p, n in zip(dalpha_pos, dalpha_neg)] according to the paper equation 8.

[jkooy]

I didn't find this at the beginning and have tried both these two. This indeed affects the update step but luckily not much on the result.