qiboteam / vqe-sun Goto Github PK

Hi @mirkoconsiglio and @igres26,

I arrived here via your preprint https://arxiv.org/abs/2106.15552

Can you me understand better if:

• all numerics is done on $L=3$ in your paper? The captions say that ED is done for $L=3$ to compare but I want to make sure I don't misunderstand that maybe larger VQEs were explored but not compared?
• the figure legends show up to 6 layers, can you comment on whether larger system sizes would be limited by training with more layers or rather by memory needed to do the simulations?

Can you comment in general what kind of system sizes were reasonable and when scaling up was becoming too costly?

Thanks in advance!