Fix a typo in cholesky_inverse documentation
#110364
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/110364
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit 30d220e with merge base 13af952 ( This comment was automatically generated by Dr. CI and updates every 15 minutes. |
torch/_torch_docs.py
Outdated
| input (Tensor): the input tensor :math:`A` of size :math:`(*, n, n)`, | ||
| consisting of symmetric positive-definite matrices | ||
| where :math:`*` is zero or more batch dimensions. | ||
| input (Tensor): input matrix :math:`u` of size :math:`(*, m, m)`, |
There was a problem hiding this comment.
why changing n into m tensor into matrix?
Please use the nomenclature used in the rest of the linalg module, like in cholesky for example.
There was a problem hiding this comment.
I used the same description as in cholesky_solve which shares exactly the same parameter u
But it is true it could be n instead of m, as cholesky uses n and not m but cholesky_solve should probably follow the same nomenclature.
There was a problem hiding this comment.
The docs for these ops are not particularly clean, as we chose not to port them to linalg. The linalg docs are properly curated, so it's best to use them as the reference in style.
If, as a follow up, you want to submit a PR writing the docs for these two ops following the styles of torch.linalg, I'd be more than happy to review it :)
There was a problem hiding this comment.
I added a commit that rewrites these two ops following what is done in linalg.cholesky. I let you have a look to the new modifications
Follow linalg.cholesky nomenclature
lezcano
added
module: docs
|
@pytorchbot merge |
Merge startedYour change will be merged once all checks pass (ETA 0-4 Hours). Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
Very small PR to fix a typo in https://pytorch.org/docs/stable/generated/torch.cholesky_inverse.html doc.
According to the current doc, the function expects$A$ , the symmetric positive-definite matrix, as input. But the examples given (and more important, the code) is using $u$ the cholesky decomposition of this matrix (like cholesky_solve).
Also, it provides a correct example of batch usage of this function.
cc @svekars @carljparker