Fix Ivy Failing Test: jax - norms.batch_norm#28920
Fix Ivy Failing Test: jax - norms.batch_norm#28920Ajay6601 wants to merge 11 commits intoivy-llc:mainfrom
Conversation
…latest changes and resolve conflicts 'main' of https://github.com/Ajay6601/ivy
all tests passed
|
|
Reformat layers : conv1d_transpose |
Sam-Armstrong
left a comment
Sam-Armstrong
left a comment
There was a problem hiding this comment.
You seem to be doing multiple things all as part of the same PR, can you change this PR to only make the necessary changes to fix the test for ivy.batch_norm, and open separate PRs for resolving any other issues? Also see my other comments. Thanks!
ivy/data_classes/array/layers.py
Outdated
| { | ||
| a: ivy.Shape(1, 56, 6), | ||
| b: ivy.Shape(1, 112, 6) | ||
| } |
There was a problem hiding this comment.
I don't understand why you've changed this to refer to ivy.Container when these methods are on ivy.Array? Can you revert you're changes to this file?
There was a problem hiding this comment.
Thank you, Sam. I have changed it.
I thought like in Ivy; there are separate implementations for:
Regular array operations (on ivy.Array)
Container operations (on ivy.Container) which apply functions to nested arrays
so when working with a single array instead of a container, you would use ivy.conv1d_transpose or ivy.static_conv1d_transpose directly on an ivy.Array . Thats why i thought like for multiple arrays we must use container. But now i got it reading through docs again.
| x: Union[tf.Tensor, tf.Variable], | ||
| filters: Union[tf.Tensor, tf.Variable], | ||
| x: Union[ivy.Array, ivy.NativeArray], | ||
| filters: Union[ivy.Array, ivy.NativeArray], |
There was a problem hiding this comment.
This was correct before.
There was a problem hiding this comment.
Oh Yeah that's correct I have been parallelly working on other issue so i might have changed this code mistakenly.
Thank you
| ret_np, ret_from_gt_np = ivy.promote_types_of_inputs(ret_np, ret_from_gt_np) | ||
| ret_dtype = str(ret_np.dtype) | ||
| ret_from_gt_dtype = str(ret_from_gt_np.dtype).replace("longlong", "int64") | ||
| assert ret_dtype == ret_from_gt_dtype, ( |
There was a problem hiding this comment.
We probably shouldn't be making any changes to this file.
Co-authored-by: Sam Armstrong <88863522+Sam-Armstrong@users.noreply.github.com>
2 participants
close #28919
close #28921