Alby M.
I completely understand your reluctance to break it up; I was more wondering from a debugging point of view.
@nw13slx looks like only some of the random generations use `generator` here? did some commits not push?
Can reproduce this with a completely different model and it only happens on PyTorch 1.10. For me it also only happens when I `torch.jit.freeze` the model.
Bumping this issue — it would be really nice to have a simple empty template so that you don't need to go through deleting old info about this repo when...
Hm right — I _think_ all such bilinearities can be expressed as a linear operation between a linear operator A and the outer product of x and y, which can...
Hi @patrick-kidger, thanks for he quick answer! This level of arcane tinkering with TorchScript definitely sounds familiar to me... :grin: The issue you link in the third bullet does make...
Hi @mhellstr , Good point--- changed on `develop`: https://github.com/mir-group/nequip/commit/da3e4bdf7ce9a25a1427f0efd15504dee71964df
Quoting the above, > The vast majority of the time this is a sign that something is wrong (such as trying to set the cutoff in the wrong distance units),...
It would be great to have this available in PyPI; it would make it much easier to use and ensure that downstream packages are getting any updates or bugfixes. It's...
> More importantly if your primary goal is to do inference, there are a couple known issue we have to address before we can have a good inference performance. One...