Alex Rogozhnikov

I miss that sometimes as well, but so far refrain from wrapping existing functionality just for the sake of this feature. Let me keep this issue open to collect cases...

@MilesCranmer this last thing is already in plans, you can read RFC #71 @cgarciae I think it makes sense to look that RFC, as to my mind it covers large...

@MilesCranmer Do you have something like Tim's zweisum in mind? https://gist.github.com/rockt/a3191f517728ea9a136a204f578d27c8 I just want to discuss some issues before they appear: parsing likely should be cached, and standard backend-guessing should...

There are several sides: - interface-wise composing and decomposing on the fly is *the right einopsy way* - speed-wise (in some frameworks) this will incur overhead for cases when rearrangement...

Option 2 looks better, but can't work with type hints, and that's a strong argument against.

this will correctly work with IDEs (pycharm / vscode), and should pass mypy (though did not check latter). There is a second part when typing matters: scripting and optimization. If...

Looking at examples ```python # your example result = einsum(batched_images, filters, "batch h w, h w channel -> batch channel") # my example result = einsum(activations, activations, "b h w...

> i.e., the first argument is a list of tensors. Lists are reserved for semantically different thing (e.g. see indexing or 'stacking' with rearrange). To weight on pattern-first and pattern-last...

and the winner is ...

> I should add tests or not Yes, tests are needed, because we need to ensure that following things are working: 1. serialization/deserialization of weights 2. serialization/deserialization/pickling/deepcopy-ing of models 3....