ivy
ivy copied to clipboard
Transpile-AI
Implemented Shuffle and Uniform for Tensorflow Frontend (Random)
Hi @saeedashrraf. Can you please review this PR? The tensorflow frontend test throws an error for test_tensorflow_uniform that I am not able to produce locally. Any advice for that?
Hey @Auxia , the error is being thrown because some of the tests generate the value None for the argument shape and tensorflow's implementation of the function does not support that. In addition to that, the argument dtype for tensorflow's implementation of the function is required to have the same dtype of minval and maxval. Keep in mind that dtype defaults to float32 and that might cause issues as well.
This PR is stale because it has been open 7 days with no activity. Remove stale label or comment or this will be closed in 7 days.
![github-actions[bot] avatar](https://avatars.githubusercontent.com/in/15368?v=4 "Published by a pub.dev verified publisher"){kind=small}
Hi @saeedashrraf, the tests are passing now for both shuffle and uniform. Can you please do a review and see if there is something I might have missed.
Hi @saeedashrraf, I am having an issue with test_uniform where it passes bfloat16 dtype to the function (which is not supported by the numpy backend). I tried to add this to the unsupported_dtypes dict, yet it still passes bfloat16. Is there a way I can either remove it from passing the said dtype entirely or atleast make it skip it for the numpy backend specifically.
I have been unable to fix the issues with uniform function but I want to get remove this branch from the main ivy repo. So I ahve removed uniform for now so that we can get this merged instead and delete this branch ASAP.
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.