01_optimisation_gd_fista convergence plot doesn't show dual objective

[hrobarts]

When we plot the convergence of PDHG in the 01_optimisation_gd_fista notebook, the dual objective is still negative after 500 iterations so it doesn't show on the log plot, but does show in the legend which is potentially confusing:

After 4500 iterations it becomes positive, so we see it on the graph and see the primal and dual objective are converging, but it's still a bit confusing that it appears out of nowhere:

Perhaps we could plot the primal and dual objective separately on a non log plot so we can see them even if they're negative? Or maybe just add to the text to explain what's happening.