DJ Rich

I'd like to cast my vote in favor of getting this functionality in. It would be nice to autodiff through xarray operations. From reading this and related threads, I'm trying...

> No one is working on __array_function__ at the moment. Implementing it has some backwards compat concerns as well, because people may be relying on np.somefunc(some_torch_tensor) to be coerced to...

> While it would be fantastic to have gpu-enabled auto-diff-able xarrays / DataArrays, an interesting development worth looking into are the named tensor in https://pytorch.org/docs/stable/named_tensor.html. This appears to be an...

> Note that your the main work in adding __array_function__ is not the dispatch mechanism, but mapping to 100% compatible APIs. That job should have gotten a lot easier now...

I've made some mild progress, but it raises a few questions. I've defined this simple Tensor subclass which meets the duck array criteria: ``` class XArrayTensor(torch.Tensor): def __new__(cls, data=None, requires_grad=False):...

Yes, it's not a package to be installed. It's just a collection of notebooks. Which notebook in particular are you trying to run?