OOM eventually when using create_graph=True with BatchL2Grad

I was trying to use my second-order optimizer [ESGD-M](https://github.com/crowsonkb/esgd) with BatchL2Grad in order to collect information on within-batch gradient variance to estimate stochastic noise (think OpenAI's gradient noise scale paper), and I kept OOMing after maybe six epochs of MNIST training. ESGD-M does a Hessian-vector product internally (not using Backpack stuff, just autograd) so it needs the user to specify create_graph=True. I assume when I use it with Backpack, something is leaking references to past computational graphs, normally these graphs are garbage collected without issue.

Thank you,
Katherine Crowson