This repository has been archived by the owner on Nov 17, 2023. It is now read-only.
fixed flaky test issue for test_operator_gpu.test_depthwise_convolution #12402
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
haojin2
approved these changes
Aug 30, 2018
lebeg
approved these changes
Aug 31, 2018
|
@mseth If it is not reproducible, why are we relaxing tolerance? |
anirudh2290
reviewed
Aug 31, 2018
| @@ -1659,7 +1658,7 @@ def test_depthwise_convolution(): | |||
| exe2.backward(exe2.outputs[0]) | |||
There was a problem hiding this comment.
can we check whether a downcast is happening when we do arr2[:] = arr1 from float64 to float32. If yes lets fix that check if the test is still flaky.
There was a problem hiding this comment.
arr1 and arr2 are both float32. Downcast is happening during populating arr1 by random normal samples. I have made the casting explicit just to avoid errors.
lebeg
added a commit
to lebeg/incubator-mxnet
that referenced
this pull request
Sep 3, 2018
…onvolution (apache#12402)" This reverts commit 58560f6.
|
@lupesko tolerance parameter is relaxed even though the issue is not reproducible as it might be the case that the exact environment is not replicated at our end, also the relaxation we are making is small. |
aaronmarkham
pushed a commit
to aaronmarkham/incubator-mxnet
that referenced
this pull request
Sep 11, 2018
…on (apache#12402) * fixed flaky test issue for test_operator_gpu.test_depthwise_convolution * Changed implicit cast to explicit cast
aaronmarkham
pushed a commit
to aaronmarkham/incubator-mxnet
that referenced
this pull request
Sep 11, 2018
…onvolution (apache#12402)" (apache#12441) This reverts commit 58560f6.
anirudh2290
pushed a commit
to anirudh2290/mxnet
that referenced
this pull request
Sep 19, 2018
…on (apache#12402) * fixed flaky test issue for test_operator_gpu.test_depthwise_convolution * Changed implicit cast to explicit cast
anirudh2290
pushed a commit
to anirudh2290/mxnet
that referenced
this pull request
Sep 19, 2018
…onvolution (apache#12402)" (apache#12441) This reverts commit 58560f6.
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
Issue not reproducible. Tolerance parameter (rtol) relaxed. This should fix the flakiness issue #12203 since mismatch percentage is small.
Checklist
Essentials
Please feel free to remove inapplicable items for your PR.
Changes
Comments