Release notes for v1.5.0-preview2 (dotnet#4787)

* Release notes for v1.5.0-preview2

* Updated automl section

* Addressed review comments

* Addressed some comments

* Updated release notes

* Added note about ORT upgrade

Author: harishsk

docs/release-notes/1.5.0-preview2/release-1.5.0-preview2.md

@@ -0,0 +1,89 @@
+# [ML.NET](http://dot.net/ml) 1.5.0-preview2
+
+## **New Features (IN-PREVIEW, please provide feedback)**
+- **TimeSeriesImputer** ([#4623](https://github.com/dotnet/machinelearning/pull/4623)) This data transformer can be used to impute missing rows in time series data.
+- **LDSVM Trainer** ([#4060](https://github.com/dotnet/machinelearning/pull/4060)) The "Local Deep SVM" usess trees as its SVM kernel to create a non-linear binary trainer. A sample can be found [here](https://github.com/dotnet/machinelearning/blob/c819d77e9250c68883713d5f1cd79b8971a11faf/docs/samples/Microsoft.ML.Samples/Dynamic/Trainers/BinaryClassification/LdSvm.cs).
+- **Onnxruntime updated to v1.2** This also includes support for GPU execution of onnx models
+- **Export-to-ONNX for below components:**
+    - SlotsDroppingTransformer ([#4562](https://github.com/dotnet/machinelearning/pull/4562))
+    - ColumnSelectingTransformer ([#4590](https://github.com/dotnet/machinelearning/pull/4590))
+    - VectorWhiteningTransformer ([#4577](https://github.com/dotnet/machinelearning/pull/4577))
+    - NaiveBayesMulticlassTrainer ([#4636](https://github.com/dotnet/machinelearning/pull/4636))
+  	- PlattCalibratorTransformer ([#4699](https://github.com/dotnet/machinelearning/pull/4699))
+  	- TokenizingByCharactersTransformer ([#4805](https://github.com/dotnet/machinelearning/pull/4805))
+  	- TextNormalizingTransformer ([#4781](https://github.com/dotnet/machinelearning/pull/4781))
+
+  
+## **Bug Fixes**
+  - Fix issue in WaiterWaiter caused by race condition ([#4829](https://github.com/dotnet/machinelearning/pull/4829))
+	- Onnx Export change to allow for running inference on multiple rows in OnnxRuntime ([#4783](https://github.com/dotnet/machinelearning/pull/4783))
+  - Data splits to default to MLContext seed when not specified ([#4764](https://github.com/dotnet/machinelearning/pull/4764))
+  - Add Seed property to MLContext and use as default for data splits ([#4775](https://github.com/dotnet/machinelearning/pull/4775))
+- **Onnx bug fixes**
+  - Updating onnxruntime version ([#4882](https://github.com/dotnet/machinelearning/pull/4882))
+  - Calculate ReduceSum row by row in ONNX model from OneVsAllTrainer ([#4904](https://github.com/dotnet/machinelearning/pull/4904))
+  - Several onnx export fixes related to KeyToValue and ValueToKey transformers ([#4900](https://github.com/dotnet/machinelearning/pull/4900), [#4866](https://github.com/dotnet/machinelearning/pull/4866), [#4841](https://github.com/dotnet/machinelearning/pull/4841), [#4889](https://github.com/dotnet/machinelearning/pull/4889), [#4878](https://github.com/dotnet/machinelearning/pull/4878), [#4797](https://github.com/dotnet/machinelearning/pull/4797))
+  - Fixes to onnx export for text related transforms ([#4891](https://github.com/dotnet/machinelearning/pull/4891), [#4813](https://github.com/dotnet/machinelearning/pull/4813))
+  - Fixed bugs in OptionalColumnTransform and ColumnSelecting ([#4887](https://github.com/dotnet/machinelearning/pull/4887), [#4815](https://github.com/dotnet/machinelearning/pull/4815))
+  - Alternate solution for ColumnConcatenatingTransformer ([#4875](https://github.com/dotnet/machinelearning/pull/4875))
+  - Added slot names support for OnnxTransformer ([#4857](https://github.com/dotnet/machinelearning/pull/4857))
+  - Fixed output schema of OnnxTransformer ([#4849](https://github.com/dotnet/machinelearning/pull/4849))
+  - Changed Binarizer node to be cast to the type of the predicted label … ([#4818](https://github.com/dotnet/machinelearning/pull/4818))
+  - Fix for OneVersusAllTrainer ([#4698](https://github.com/dotnet/machinelearning/pull/4698)) 
+  - Enable OnnxTransformer to accept KeyDataViewTypes as if they were UInt32 ([#4824](https://github.com/dotnet/machinelearning/pull/4824))
+  - Fix off by 1 error with the cats_int64s attribute for the OneHotEncoder ONNX operator ([#4827](https://github.com/dotnet/machinelearning/pull/4827))
+  - Changed Binarizer node to be cast to the type of the predicted label … ([#4818](https://github.com/dotnet/machinelearning/pull/4818))
+  - Updated handling of missing values with LightGBM, and added ability to use (0) as missing value ([#4695](https://github.com/dotnet/machinelearning/pull/4695))
+  - Double cast to float for some onnx estimators  ([#4745](https://github.com/dotnet/machinelearning/pull/4745))
+  - Fix onnx output name for GcnTransform ([#4786](https://github.com/dotnet/machinelearning/pull/4786))
+- Added support to run PFI on uncalibrated binary classification models ([#4587](https://github.com/dotnet/machinelearning/pull/4587))
+- Fix bug in WordBagEstimator when training on empty data ([#4696](https://github.com/dotnet/machinelearning/pull/4696))
+- Added Cancellation mechanism to Image Classification (through the experimental nuget) (fixes #4632) ([#4650](https://github.com/dotnet/machinelearning/pull/4650))
+-	Changed F1 score to return 0 instead of NaN when Precision + Recall is 0 ([#4674](https://github.com/dotnet/machinelearning/pull/4674))
+-	TextLoader, BinaryLoader and SvmLightLoader now check the existence of the input file before training ([#4665](https://github.com/dotnet/machinelearning/pull/4665))
+- ImageLoadingTransformer now checks the existence of input folder before training ([#4691]( https://github.com/dotnet/machinelearning/pull/4691))
+- Use random file name for AutoML experiment folder ([#4657](https://github.com/dotnet/machinelearning/pull/4657))
+- Using invariance culture when converting to string  ([#4635](https://github.com/dotnet/machinelearning/pull/4635))
+- Fix NullReferenceException when it comes to Recommendation in AutoML and CodeGenerator ([#4774](https://github.com/dotnet/machinelearning/pull/4774))
+
+## **Enhancements**
+- Added in support for System.DateTime type for the DateTimeTransformer ([#4661](https://github.com/dotnet/machinelearning/pull/4661))
+- Additional changes to ExpressionTransformer ([#4614](https://github.com/dotnet/machinelearning/pull/4614))
+- Optimize generic MethodInfo for Func<TResult> ([#4588](https://github.com/dotnet/machinelearning/pull/4588))
+- Data splits to default to MLContext seed when not specified ([#4764](https://github.com/dotnet/machinelearning/pull/4764))
+- Added in DateTime type support for TimeSeriesImputer ([#4812](https://github.com/dotnet/machinelearning/pull/4812))
+
+## **Test updates**
+- **Code analysis updates**
+  - Update analyzer test library ([#4740](https://github.com/dotnet/machinelearning/pull/4740))
+  - Enable the internal code analyzer for test projects ([#4731](https://github.com/dotnet/machinelearning/pull/4731))
+  - Implement MSML_ExtendBaseTestClass (Test classes should be derived from BaseTestClass) ([#4746](https://github.com/dotnet/machinelearning/pull/4746))
+  - Enable MSML_TypeParamName for the full solution ([#4762](https://github.com/dotnet/machinelearning/pull/4762))
+  - Enable MSML_ParameterLocalVarName for the full solution ([#4833](https://github.com/dotnet/machinelearning/pull/4833))
+  - Enable MSML_SingleVariableDeclaration for the full solution ([#4765](https://github.com/dotnet/machinelearning/pull/4765))
+- **Better logging from tests**
+  - Ensure tests capture the full log ([#4710](https://github.com/dotnet/machinelearning/pull/4710))
+  - Fix failure to capture test failures ([#4716](https://github.com/dotnet/machinelearning/pull/4716))
+  - Collect crash dump upload dump and pdb to artifact ([#4666](https://github.com/dotnet/machinelearning/pull/4666))
+- Enable Conditional Numerical Reproducibility for tests ([#4569](https://github.com/dotnet/machinelearning/pull/4569))
+- Changed all MLContext creation to include a fixed seed ([#4736](https://github.com/dotnet/machinelearning/pull/4736))
+- Fix incorrect SynchronizationContext use in TestSweeper ([#4779](https://github.com/dotnet/machinelearning/pull/4779))
+
+## **Documentation Updates**
+- Update cookbook to latest API ([#4706](https://github.com/dotnet/machinelearning/pull/4706))
+- Update documentation to stop mentioning interfaces that no longer exist ([#4673](https://github.com/dotnet/machinelearning/pull/4673))
+- Roadmap update ([#4704](https://github.com/dotnet/machinelearning/pull/4704))
+- Added release process documentation to README.md ([#4402](https://github.com/dotnet/machinelearning/pull/4402))
+- Fix documentation of SvmLightLoader ([#4616](https://github.com/dotnet/machinelearning/pull/4616))
+- Correct KMeans scoring function doc ([#4705](https://github.com/dotnet/machinelearning/pull/4705))
+- Several typo fixes thanks to [@MaherJendoubi](https://github.com/MaherJendoubi) ([#4627](https://github.com/dotnet/machinelearning/pull/4627), [#4631](https://github.com/dotnet/machinelearning/pull/4631), [#4626](https://github.com/dotnet/machinelearning/pull/4626) [#4617](https://github.com/dotnet/machinelearning/pull/4617), [#4633](https://github.com/dotnet/machinelearning/pull/4633), [#4629](https://github.com/dotnet/machinelearning/pull/4629), [#4642](https://github.com/dotnet/machinelearning/pull/4642))
+- Other typo fixes: ([#4628](https://github.com/dotnet/machinelearning/pull/4628), [#4685](https://github.com/dotnet/machinelearning/pull/4685), [#4885](https://github.com/dotnet/machinelearning/pull/4885))
+
+
+## **Breaking Changes**
+- None
+
+
+
+
+