frontends.torch.linalg: add support for cross #7656
Conversation
ivy-leaves
added
the
PyTorch Frontend
There was a problem hiding this comment.
Hi, thanks for submitting a PR. With this task, your goal is to recreate the Torch functions using functions from the Ivy framework. It is possible that you will have to use a combination of Ivy functions in order to recreate this behaviour :)
More info here
| @to_ivy_arrays_and_back | ||
| @with_unsupported_dtypes({"1.11.0 and below": ("float16",)}, "torch") | ||
| def cross(input, other, *, dim=-1, out=None): | ||
| return torch_frontend.cross(input, other, dim, out=out) |
There was a problem hiding this comment.
You would need to use Ivy functions here, rather than torch functions.
There was a problem hiding this comment.
Hi,
Thanks for the review. I have implemented this function according to the instructions provided in the ToDo list here:
Note: If the Linear Algebra function to be implemented has identical behavior to another PyTorch function, you should simply keep an alias in the linalg.py file rather than creating a duplicate implementation.
Here, it is explicitly mentioned not to provide a duplicate implementation of an already implemented PyTorch function that has an identical behavior, in my case that is torch.cross, but just use it to define an alias in ../torch/linalg.py as I've done so. The example provided in that ToDo list highlights exactly what I've implemented. Can you kindly cross verify with the guidelines we were provided and what I've implemented? 🙂 Thanks
There was a problem hiding this comment.
I have checked and this is the correct protocol :)
|
@jkeane508 The conflicts have been resolved :) |
Close #7597