Merge master into release/preview for 0.9 #2004
Merged
Conversation
This file contains hidden or 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
Added VarVector static extension for OneHotHashEncoding
* adding new ci leg * removing . * Fixing and disabling tests for netcoreapp3.0 * skipping some more tests and name change * added comments and conditional fact * name changed to NotNetCore30 * name changed and reding cli version from the file * name changed to DotnetCLIVersion.netcoreapp.latest.txt * contracts assert cprrected for avx=intrinscis fixing new tests broken by other commits Fixing Tests
…ts (#1777) * correcting Documentation, adding asserts and disabling failing ci tests * Correcting asserts
* Move evaluation results to stand alone metrics classes
…1775) * Prevent models from copying from each leg * Copying now checks for condition * Add correct property * Minor change to condition logic
* Adding LDA sample to Microsoft.ML.Samples * review comments - 1. nit changes
* Remove IRandom and replace with System.Random. * Make TauswortheHybrid internal/BestFriend. * Add NextSingle and NextSigned back. * Override Next(int, int) and NextBytes since they are public. * Change Single and Double to float and double.
Updating Binary Classification XML Docs to the latest version of the PFI documentation text.
* Remove ISchematized interface from the codebase. 1. Remove ISchematized 2. For any class that requires a Schema, we add an Schema as its field 3. Rename Schema to OutputSchema in IRowToRowMapper * Address comments * Clean redundant Schema field and point Schema to OutputSchema * Replace Schema in SingleValueRowMapper with OutputSchema
* Add release notes for 0.8
* update transform to use onnxruntime library instead of sonoma * Updated to conform to newer API * Resolved PR comments * minor fix * updated per PR comments * Resolved more PR comments * added functionality to pull input/output names from model, without user enumerating them explicitly * update version to mlscoring 0.1.5
* Fix CodeGen command and add a unit test * Disable new unit test on Linux and Mac * PR comments * Address PR comments * Fix generated code template * Update baseline * Update baseline
…ator with new data (#1805)
* Stop using IColumn in predicted label scorers. * Stop using IColumn in static schema shape analysis. * Stop using IColumn in linear model statistics creation. * Stop using IColumn in FastTree statistics creation. * Stop using IColumn in benchmarking. * Stop using IColumn in many tests. * Add minor conveniences to metadata builder. * Allow metadata builder to have metadata of metadata. * Add appropriate validation in certain places. * Put warnings on an inappropriate method dealing with Batch that should not exist. * Remove IColumn.
…eful prediction engine. (#1762) * Sample for IID spike and changepoint detection using time series stateful prediction engine.
* Added SsaChangePointDetectorPrediction sample. * Completed SsaChangePointDetectorPrediction sample. * Added SsaSpikeDetectorPrediction sample. * Cleanup usings, etc. * Added note when saved and loaded from disk.
* Add ScaleBy test * Add VBuffer unit tests, and fix a couple of bugs * Fix test tolerance * Address PR comments and fix unit test tolerance * Fix AlignedArray bug
* Removing Aligned Array usage from rff and timeseries. Removing aligned matrix and cpumathaligned entirely. moving Aligned array to FactorAware where it is only being used * adding some asserts * unrolling the loop, aligned removed from name, using new float comparision * enabling some more tests * case comment added, tanners feedback
… classes. (#1814) * Slot cursor now an abstract class and not an ICursor. * Members of ICounted moved to IRow. * Members of ICursor moved to IRowCursor. * Remove ICounted and ICursor. * Remove IRowCursor from places that previously had to accomodate different sorts of cursors. * Add benchmark for cache data view cursor/seeker access. * Rename private RowCursor implementations Cursor. * IRowCursor is now an abstract class. * Rename IRowCursor to RowCursor, as well as a few other similar classes. * IRow is now an abstract class. * Rename private IRow implementations named Row, RowImpl. * Rename IRow to Row, and IStatefulRow to StatefulRow.
* Rename KMeansPredictor to KMeansModelParameters, rename namespace, and internalize interfaces except ICanSaveModel * Internalizing SaveCore * Adding sample * BestFriend attribute for the correct method overridden by LightGbmBinaryPredictor.SaveCore * Making FastTree best friends with LightGBM * Update core_ep-list.tsv * Reverting libmf commit id * Explicitly implementing IValueMapper and ICanSaveInTextFormat everywhere * Final cleanup * addressing some comments * reverting namespace change * update core_ep-list.tsv
* Make SchemaShape.Column a struct instead of a class * Make SchemaShape an IReadOnlyList<SchemaShape.Column> and remove SchemaException * Make a (best) friend * Address comments * Make all member functions of SchemaShape.Column be best friends * Address comments * Change Columns' type from Column[] to ImmutableList<Column> * Fix build * Remove unnecessary checks because for some nullable columns * Address comments * SchemaShape is only a list not a class containing a list * Use array for immutable field and add a best friend function * Assert doesn't need variable name
* Remove auto-cache mechanism * Add caching usage into a sample and tests
This reverts commit 72ec121. A bug was detected that resulted in non-deterministic calculation, since the underlying C++ code was written in a way apparently that required alignment to produce consistent results, so of course just removing the alignment and calling an only slightly modified algorithm compromised determinism, resulting in test failure for RFF in particular.
* Row now disposable, and implements recommended disposable pattern. * Internalize IRowMapper and its implementing methods. * Have IRowToRowMapper return disposable Rows and no longer have disposer delegate.
* Adding a binary classification PFI Example, breaking the PFI examples into different files in a subfolder, and correcting XMLDocs links.
* Have RoleMappedSchema use Schema.Column not ColumnInfo. * Have EvaluatorUtils and calling code exploit Schema.Column not ColumnInfo. * Replace other few miscellaneous usages of ColumnInfo with Schema.Column. * Remove ColumnInfo. * Improve error reporting uniformity through usage of SchemaMismatch contracts helpers in several places.
* Added rffBenchmark * scikit removed and another reference added * Converting to api * adding cacheCheck point to correct place * comment added about dataset
* moving reflection-based read out of the legacy project. * adding more members to LoadColumnAttribute * Breaking down the ordinal/range into ordinal, start, end, columnIndices. Introducing test.
* Addition of the ValueMappingEstimator and ValueMappingTransform. This will be replacing the TermLookupTransform and provide a way to specify the mapping betweeen two values (note this is specified and not trained). A user can specify the mapping by providing a keys list and values list that must be equal in size. The Estimator will then generate a 1-1 mapping based on the two lists. The PR references #754 which covers the conversion of Transformer to use the new Estimator API.
* Reducing public surface of remaining predictors * Addressing comments, adding doc strings * Addressing comments, adding sample
…ed namespaces from source files. (#1963) * Sort namespaces as stylecop rules/.net convention and remove unused namespaces from source files.
* use channel info instead of printf
* Support back compat
* loads inf/nan properly * Address comments
# Conflicts: # docs/samples/Microsoft.ML.Samples/Dynamic/IidChangePointDetectorTransform.cs # docs/samples/Microsoft.ML.Samples/Dynamic/IidSpikeDetectorTransform.cs # docs/samples/Microsoft.ML.Samples/Dynamic/LdaTransform.cs # docs/samples/Microsoft.ML.Samples/Dynamic/SsaChangePointDetectorTransform.cs # docs/samples/Microsoft.ML.Samples/Dynamic/SsaSpikeDetectorTransform.cs # docs/samples/Microsoft.ML.Samples/Microsoft.ML.Samples.csproj # docs/samples/Microsoft.ML.Samples/Program.cs # src/Microsoft.ML.Data/Scorers/FeatureContributionCalculation.cs # src/Microsoft.ML.ImageAnalytics/EntryPoints/ImageAnalytics.cs # src/Microsoft.ML.Transforms/PermutationFeatureImportanceExtensions.cs
Contributor
Author
|
No Squash merge. |
shauheen
requested review from
Ivanidzo4ka,
TomFinley,
codemzs,
eerhardt and
sfilipi
codemzs
approved these changes
Jan 3, 2019
eerhardt
approved these changes
Jan 3, 2019
ghost
locked as resolved and limited conversation to collaborators
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
23 participants
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.
This PR merges master into release branch for 0.9