Hi, the bandstructure plots also contain bands that are not fully converged (if I understand n_bands_converge in AdaptiveBands correctly). This can be misleading if you expect all plotted bands to be converged.
To give some background:
I encountered the following issue with Wannierization, where the disentanglement step would fail to select the right band close to the Gamma point (see the red line). It turns out that I was only passing the 15 first bands to Wannier, but the expected band is the 16th. I assumed that the plot was only showing the first 15 bands because I set n_bands_converge=15.
Would it be a good idea to plot non-converged bands a bit differently? (For example with a dotted line).
Hi, the bandstructure plots also contain bands that are not fully converged (if I understand
n_bands_convergeinAdaptiveBandscorrectly). This can be misleading if you expect all plotted bands to be converged.To give some background:
I encountered the following issue with Wannierization, where the disentanglement step would fail to select the right band close to the Gamma point (see the red line). It turns out that I was only passing the 15 first bands to Wannier, but the expected band is the 16th. I assumed that the plot was only showing the first 15 bands because I set
n_bands_converge=15.Would it be a good idea to plot non-converged bands a bit differently? (For example with a dotted line).