Add hstack, vstack, and bmat to cupyx.scipy.sparse#2665
Add hstack, vstack, and bmat to cupyx.scipy.sparse#2665niboshi merged 29 commits intocupy:masterfrom
hstack, vstack, and bmat to cupyx.scipy.sparse#2665Conversation
|
Nice start, @cjnolet. I left a few initial comments, but haven't done a close review |
|
@niboshi, I believe this PR is ready for review. |
|
@niboshi, Thank you for the review. I've implemented your review feedback. I believe this is ready for re-review. |
|
What's the status with this? I could use this functionality as well, and I'm happy to help get the feature finalized. |
|
@niboshi, checking in. We really need this feature in RAPIDS. Do you have any further change requests on this PR? |
niboshi
left a comment
There was a problem hiding this comment.
I'm sorry for delay. Please check.
Co-Authored-By: niboshi <niboshi000@gmail.com>
|
@niboshi requested changes have been made |
|
@niboshi requested changes have been made |
|
Thanks! |
|
Successfully created a job for commit 997bc57: |
|
Jenkins CI test (for commit 997bc57, target branch master) failed with status FAILURE. |
|
@niboshi the requested changes have been made. I used |
|
Thanks! Why would you need to add |
|
I can use decimals if you prefer. I guess I'm just in the habit of using single precision float wherever possible because of the performance implications of double precision. Let me know if you'd like me to change it and I'll do so quickly. |
|
@niboshi, I updated it anyways. Makes the docs look cleaner. |
|
We don't use |
This comment has been minimized.
This comment has been minimized.
|
Forgive me, @niboshi, I should have regenerated the output for the example after modifying it. |
|
NP, I didn't notice either. 😅 |
|
Successfully created a job for commit 171ff88: |
|
Jenkins CI test (for commit 171ff88, target branch master) succeeded! |
hstack, vstack, and bmat to cupyx.scipy.sparse
|
Thank you so much! |
|
Thank you, @niboshi! |
I started with the corresponding functions from
scipy.sparseand made some modifications to make them work withcupy. Unfortunately, sincecupy.sparse.coo_matrixdoes not take a dense matrix as an argument to its constructor, the example in the pydocs will need to change slightly.I will be working to add tests next.