[Predator] Fix weight flaky loglinear training test.

[Predator] Fix weight flaky loglinear training test.

BUG=701238
TBR=stgao@chromium.org

Review-Url: https://codereview.chromium.org/2753503002
Committed: https://chromium.googlesource.com/infra/infra/+/68e0dc8cc731ac339d136215e7f684a82a73ccc6

[Sharu Jiang, 2017-03-14 03:55:29]

Description was changed from

==========
[Predator] Fix weight dimension bug.

BUG=701238
==========

to

==========
[Predator] Fix weight dimension bug.

BUG=701238
TBR=stgao@chromium.org
==========

[Sharu Jiang, 2017-03-14 03:55:29]

katesonia@chromium.org changed reviewers:
+ stgao@chromium.org

[Sharu Jiang, 2017-03-14 03:56:36]

Commit this simple fix first to prevent flaky tests.

[Sharu Jiang, 2017-03-14 03:56:47]

The CQ bit was checked by katesonia@chromium.org

[commit-bot: I haz the power, 2017-03-14 03:56:55]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2017-03-14 04:05:35]

CQ is committing da patch.

Bot data: {"patchset_id": 1, "attempt_start_ts": 1489463807456180, "parent_rev":
"bcf983131a051a03275bd245c07ca4cb3d28eb06", "commit_rev":
"68e0dc8cc731ac339d136215e7f684a82a73ccc6"}

[commit-bot: I haz the power, 2017-03-14 04:05:38]

Description was changed from

==========
[Predator] Fix weight dimension bug.

BUG=701238
TBR=stgao@chromium.org
==========

to

==========
[Predator] Fix weight dimension bug.

BUG=701238
TBR=stgao@chromium.org

Review-Url: https://codereview.chromium.org/2753503002
Committed:
https://chromium.googlesource.com/infra/infra/+/68e0dc8cc731ac339d136215e7f68...
==========

[commit-bot: I haz the power, 2017-03-14 04:05:39]

Committed patchset #1 (id:1) as
https://chromium.googlesource.com/infra/infra/+/68e0dc8cc731ac339d136215e7f68...

[stgao, 2017-03-14 04:34:33]

code lgtm, but I still can't understand the flakiness with the fixing.

https://codereview.chromium.org/2753503002/diff/1/appengine/findit/crash/logl...
File appengine/findit/crash/loglinear/test/weight_test.py (left):

https://codereview.chromium.org/2753503002/diff/1/appengine/findit/crash/logl...
appengine/findit/crash/loglinear/test/weight_test.py:63: with
self.assertRaisesRegexp(Exception,
So you mentioned in the bug "when the trained weights have zeros, assertion
exceptions will raise". But in this testcase, do we use randomly-generated
trained weights? If not, I'm wondering how this became flaky? Is it because of
precision issue in floating point wen compared to zero?

[Sharu Jiang, 2017-03-14 05:06:13]

https://codereview.chromium.org/2753503002/diff/1/appengine/findit/crash/logl...
File appengine/findit/crash/loglinear/weight.py (left):

https://codereview.chromium.org/2753503002/diff/1/appengine/findit/crash/logl...
appengine/findit/crash/loglinear/weight.py:85: assert len(self) ==
len(meta_feature), Exception(
This assertion should be deleted. Because the ``DropZeroWeights`` will drop all
zeros in meta weight, which will make the length of the ``meta_weight``
different from the length of ``meta_feature``.

In
https://cs.chromium.org/chromium/infra/appengine/findit/crash/loglinear/test/...,
line 55.

If the trained weight happens to be 0 (the training set is randomly generated,
so the training results vary every time), the length of meta weight would be 0,
however the length of meta feature is still 1, the assertion will fail.

[Sharu Jiang, 2017-03-14 05:09:17]

https://codereview.chromium.org/2753503002/diff/1/appengine/findit/crash/logl...
File appengine/findit/crash/loglinear/test/weight_test.py (left):

https://codereview.chromium.org/2753503002/diff/1/appengine/findit/crash/logl...
appengine/findit/crash/loglinear/test/weight_test.py:63: with
self.assertRaisesRegexp(Exception,
On 2017/03/14 04:34:33, stgao (catching up) wrote:
> So you mentioned in the bug "when the trained weights have zeros, assertion
> exceptions will raise". But in this testcase, do we use randomly-generated
> trained weights? If not, I'm wondering how this became flaky? Is it because of
> precision issue in floating point wen compared to zero?

No, the flaky test is ``testTrainWeights`` in training_test.py. It is because
the training data set of loglinear model is randomly generated.

[Sharu Jiang, 2017-03-14 05:12:34]

Description was changed from

==========
[Predator] Fix weight dimension bug.

BUG=701238
TBR=stgao@chromium.org

Review-Url: https://codereview.chromium.org/2753503002
Committed:
https://chromium.googlesource.com/infra/infra/+/68e0dc8cc731ac339d136215e7f68...
==========

to

==========
[Predator] Fix weight flaky loglinear training test.

BUG=701238
TBR=stgao@chromium.org

Review-Url: https://codereview.chromium.org/2753503002
Committed:
https://chromium.googlesource.com/infra/infra/+/68e0dc8cc731ac339d136215e7f68...
==========