Skip to content
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

Modified Tensorflow ImageClassification tests to make them faster #4447

Draft
wants to merge 3 commits into
base: master
from

Conversation

@harshithapv
Copy link
Member

harshithapv commented Nov 6, 2019

  1. Added a IClassFixture to create a workspace path only once before all the tests are run.
  2. Allow re-use of cached values wherever possible.
    Speed-ups:
    Before Tensorflow Test fixes - Tests took 8.67 minutes on an average.
    After Tensorflow Test fixes - Tests took 7.33 minutes on an average.
    These are results with Polynomial LR test disabled. We expect better speed ups as we add more tests.
harshithapv added 2 commits Nov 6, 2019
…ure to allow for creating a workspace path exactly once for all the tests.
@codecov

This comment has been minimized.

Copy link

codecov bot commented Nov 6, 2019

Codecov Report

Merging #4447 into master will decrease coverage by 15.4%.
The diff coverage is 82.05%.

@@             Coverage Diff             @@
##           master    #4447       +/-   ##
===========================================
- Coverage   90.11%   74.71%   -15.41%     
===========================================
  Files         200      906      +706     
  Lines       36848   159298   +122450     
  Branches     1910    17149    +15239     
===========================================
+ Hits        33206   119019    +85813     
- Misses       3337    35465    +32128     
- Partials      305     4814     +4509
Flag Coverage Δ
#Debug 74.71% <82.05%> (-15.41%) ⬇️
#production 70.09% <ø> (?)
#test 90.07% <82.05%> (-0.05%) ⬇️
Impacted Files Coverage Δ
...cenariosWithDirectInstantiation/TensorflowTests.cs 90.03% <82.05%> (-1.12%) ⬇️
src/Microsoft.ML.Data/Model/Pfa/ModelUtils.cs 0% <0%> (ø)
...ML.Data/Transforms/ColumnConcatenatingEstimator.cs 83.58% <0%> (ø)
src/Microsoft.ML.Data/Model/Pfa/PfaUtils.cs 0% <0%> (ø)
...t.ML.Data/DataLoadSave/CompositeLoaderEstimator.cs 80% <0%> (ø)
...t.ML.Data/DataLoadSave/Text/LoadColumnAttribute.cs 100% <0%> (ø)
...ft.ML.Data/DataLoadSave/Database/DatabaseLoader.cs 52.66% <0%> (ø)
src/Microsoft.ML.Core/Data/ModelSaveContext.cs 91.57% <0%> (ø)
...er/Multiclass/EnsembleMulticlassModelParameters.cs 84.4% <0%> (ø)
... and 698 more
@harshithapv harshithapv changed the title Modified Tensoflow Imageclassification tests to make them faster Modified Tensorflow ImageClassification tests to make them faster Nov 6, 2019
@harshithapv harshithapv requested review from codemzs, ashbhandare and bpstark Nov 7, 2019
@codemzs

This comment has been minimized.

Copy link
Member

codemzs commented Nov 8, 2019

@harshithapv Can we please get before and after times to understand the speed ups? I understand there will more speed ups when LR polynomial tests go in so we can perhaps wait for that test to go in.

@harshithapv

This comment has been minimized.

Copy link
Member Author

harshithapv commented Nov 8, 2019

@harshithapv Can we please get before and after times to understand the speed ups? I understand there will more speed ups when LR polynomial tests go in so we can perhaps wait for that test to go in.

Added Speed ups in the PR description.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
2 participants
You can’t perform that action at this time.