Maksim Levental

>https://github.com/OmixVisualization/qtjambi/blob/master/www/How-to-bundle-Qt-libraries.md I saw that but didn't consider that would be an issue - was imagining you'd just expect people to have Qt installed but it makes sense to distribute. >If...

@seldridge thanks for the quick response. i'll try to rollback to 1.4.2 to get this to work (though i'm betting that'll break something else lol) but looking at https://github.com/chipsalliance/firrtl/issues/2203 am...

> Thanks for the contribution! Yes, please add a correctness test to make sure the old and new DP are numerically the same. @zhuohan123 Easiest place to put the correctness...

i'm still getting this segfault on ubuntu 18. even with the imagemagick bound it doesn't work. EKami's solution of softlinking works

> > Not sure if this is useful for anyone, but I thought I'd post my workaround here in the meantime. Basically the same as the original post, except that...

>Correct. I suspect that the OPT model needs to be wrapped in Python to just return the logits or something. The architecture of the model does not easily lend itself...

> Can you just do `return self.forward()[0]` on the result of the model? Or `self.forward()[0][0]` or something. ``` %852 = torch.prim.TupleConstruct %134, %139 : !torch.tensor, !torch.tensor -> !torch.tuple loc(#loc0) %853...

It's probably this same scalar promotion bug that just got patched in upstream pytorch https://github.com/pytorch/pytorch/issues/74400#event-6574874903

@silvasean I would assume so - that bug was about python scalars getting promoted to tensors (and then having any record of that promotion erased at the python level during...

> > python scalars go in, python scalars come out > > So `torch.ops.aten.ceil(1.5)` should expect scalar output rather than Tensor right? @cathyzhyi [this is the PR](https://github.com/pytorch/pytorch/pull/76835) that fixed the...