-
Notifications
You must be signed in to change notification settings - Fork 5.7k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
feat: added quantile function (new) #23580
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
PR Compliance Checks Passed!
@lucasalavapena - Please review this |
If you are working on an open task, please edit the PR description to link to the issue you've created. For more information, please check ToDo List Issues Guide. Thank you 🤗 |
@lucasalavapena This is the infor from Github-bot on many PRs so I want to highlight as it says we can "ignore the run_tests (1), run_tests (2), … jobs, and only look at the display_test_results job. 👀" which means it is good to go. Please let me know. `Thanks for contributing to Ivy! 😊👏
|
You have to click on the details and actually look at the results. They are still failing... |
@jaskiratsingh2000 hi thank you for mentioning the bot, the idea of that message meant to go directly to combined tests instead of checking each of them which we updated the message on last bot and removed it though. PLus, could you add a closing word to the PR reference? it would help to connect it with the issue and auto-close it, |
You are linking the wrong issue? That is for backends. You are adding a frontend. Plus you should link the quantile |
@lucasalavapena I actually tried but unabel to resolve them at my end. I did actually what I could do during debugs. Would appreciate if locally you cancpush them up once? |
@lucasalavapena Apology in that part here is the master issue - #15045 |
All I did was make the |
@lucasalavapena you mean where I refer to |
No, I explained this better in the other PR in which I showed you part of the paddle paddle code. Your |
Thanks for letting me know 😄 . I updated your first PR comment so it should be all good, please do that next time or update any existing PR you may have accordingly! |
@lucasalavapena I made the |
@lucasalavapena or the another opotion that can be done is
|
@lucasalavapena I tried adding what I mentioned to you on Discord. but there is no luck in surpassing those checks. I'm not able to debug this further. Therefore, I have also raised the concern on Discord Testing channel but there have been no response from them so far. Please, would really appreciate if you could over-commit on my previous or let me know what change I should made it now? |
@lucasalavapena As you mentioned that you were able to make a change and run the test case can you just tell me where exactly made the change? I'm struggling with that even after debugging. and I feel this is at the last stage. Please let me know if you can this would really means a lot. |
@Ishticode @lucasalavapena Please please let me know on this. |
@lucasalavapena I'm getting this error which I'm unable to understand thouh I know it is an issue of module imports but how to solve it?
|
input_dtypes, (x,), axis = dtype_x_and_axis | ||
# Convert axis if it's a tuple | ||
if isinstance(axis, tuple): | ||
axis = axis[0] if len(axis) == 1 else list(axis) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
what's the purpose of this logic, I don't think this is required.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Since axis varibale has to be int not tuples when accepting parameters so that is wy I added it.
) | ||
@to_ivy_arrays_and_back | ||
def quantile(x, q, axis=None, keepdim=False): | ||
return ivy.quantile(x, q, axis=axis, keepdims=keepdim) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Paddle quantile returns float64 to obtain higher precision.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
So should I mention only float64?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
It's not fixed "AssertionError: returned dtype = float32, ground-truth returned dtype = float64"
this is not related to the supported dtypes but the return dtype of the function
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@Sarvesh-Kesharwani This has to be implemented within the Test file or this one?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
@Sarvesh-Kesharwani I have added. Can you check?
you need to fix the lint errors as well. |
@Sarvesh-Kesharwani lint error is resolved and the changes you requested I made. Still I feel these tests cases fail. Not sure why and how to resolve them |
Plus there is a shape mismatch error as well, you need to update your function logic to return the shape which matches the ground_truth shape |
@Sarvesh-Kesharwani could you please let me know that error seems to be within the test file or another file? Looking forward to hearing from you. Thanks, and Regards |
@Sarvesh-Kesharwani I have resolve the both the errors which you mentioned to nme. Can you please review it again? |
This PR has been labelled as stale because it has been inactive for more than 7 days. If you would like to continue working on this PR, then please add another comment or this PR will be closed in 7 days. |
This PR has been closed because it has been marked as stale for more than 7 days with no activity. |
Close #19176