Open jenninglim opened 2 years ago
Seems like a composite compliance issue @zou3519
I thought that functionalization should be able to fix this, but seems to segfault. Any ideas @bdhirsh?
I have a local fix for jacfwd(functionalize(f))
, but it doesn't fix the problem.
Is this a composite compliance issue? It looks like matrix_exp
isn't a composite op (native function entry).
Oh hmm, it looks like it's because the derivative formula for matrix_exp
isn't "composite compliant" (although in this case, the derivative formula isn't an op, it's just a function. Defined here, where it uses a bunch of view + inplace ops)
I wonder if there's a reasonable way to get functionalization to interpose in the backward
Why shouldn't the backwards pass be functionalizable?
I am trying to calculate the jacobian of a matrix exponential. However, it yields the following error
Is there a fix for this?
The error can be reproduced by the following:
Thanks.