Something changed in `pytensor > 2.12.3` (and thus `pymc > 5.6.1`) that makes my `pytensor.gradient.grad` call get stuck - any ideas?

Questions v5
17

In the context of that source code, 1D are standard multivariate transforms of vector to vector functions. 0D is for elemwise/scalars transforms.

So don’t be confused by the 0/1/2. It sounds like you want to handle batched vector transforms like @jessegrabowski mentioned.

Can we have details on the exact type of transformation that’s going on?

1 Like