Upgrade TF.Net version from 0.10.10 to 0.11.3 #4205

ashbhandare · 2019-09-11T21:43:23Z

Fixes #4204

Includes a few API changes necessary to upgrade TF.Net version.
Verified that ML.Samples and Tests run.

…decode_image

…NET0.11.2 to reproduce memory corruption.

dnfclas · 2019-09-11T21:43:38Z

All CLA requirements met. #Resolved

codemzs · 2019-09-11T21:48:26Z

Why are the below files added? They should not be added, please remove them.

#Resolved

codemzs · 2019-09-11T21:48:54Z

docs/samples/Microsoft.ML.Samples/Dynamic/TensorFlow/ImageClassification.cs

@@ -34,11 +34,13 @@ public static void Example()
            var idv = mlContext.Data.LoadFromEnumerable(data);

            // Create a ML pipeline.
+
            var pipeline = mlContext.Model.LoadTensorFlowModel(modelLocation)


remove this new line. #Resolved

codemzs · 2019-09-11T21:49:08Z

docs/samples/Microsoft.ML.Samples/Dynamic/TensorFlow/ImageClassification.cs

            var pipeline = mlContext.Model.LoadTensorFlowModel(modelLocation)
                .ScoreTensorFlowModel(
                new[] { nameof(OutputScores.output) },
                new[] { nameof(TensorData.input) }, addBatchDimensionInput: true);
-
+
+


remove this line and please revert this file. #Resolved

codemzs · 2019-09-11T21:50:05Z

src/Microsoft.ML.Dnn/DnnRetrainTransform.cs

@@ -245,13 +245,14 @@ private void CheckTrainingParameters(DnnRetrainEstimator.Options options)

                tfInputShape = new TensorShape(colTypeDims);
            }
-            if (tfInputShape.NDim != -1)
+            if (tfInputShape.ndim != -1)


tfInputShape.ndim [](start = 16, length = 17)

cache this value. #Resolved

codemzs · 2019-09-11T21:50:33Z

src/Microsoft.ML.Dnn/DnnRetrainTransform.cs

-            loss = tensor.Length > 0 && tensor[0] != IntPtr.Zero ? (float)tensor[0].Data<float>()[0] : 0.0f;
-            metric = tensor.Length > 1 && tensor[1] != IntPtr.Zero ? (float)tensor[1].Data<float>()[0] : 0.0f;
-            var b = tensor.Length > 2 && tensor[2] != IntPtr.Zero ? (float[])tensor[2].Data<float>() : null;
+            //var buffer = tensor[0].Data();


//var buffer = tensor[0].Data(); [](start = 12, length = 32)

delete. #Resolved

bpstark · 2019-09-11T21:47:50Z

src/Microsoft.ML.Dnn/DnnRetrainTransform.cs

+                var newShape = new int[tfInputShape.ndim];
+                var dims = tfInputShape.dims;
+                newShape[0] = dims[0] == 0 || dims[0] == -1 ? batchSize : dims[0];
+                //var v = tfInputShape[0].Equals(0);


should remove unused code rather than commenting it out. #Resolved

bpstark · 2019-09-11T21:48:10Z

src/Microsoft.ML.Dnn/DnnRetrainTransform.cs

-            loss = tensor.Length > 0 && tensor[0] != IntPtr.Zero ? (float)tensor[0].Data<float>()[0] : 0.0f;
-            metric = tensor.Length > 1 && tensor[1] != IntPtr.Zero ? (float)tensor[1].Data<float>()[0] : 0.0f;
-            var b = tensor.Length > 2 && tensor[2] != IntPtr.Zero ? (float[])tensor[2].Data<float>() : null;
+            //var buffer = tensor[0].Data();


same here #Resolved

codemzs · 2019-09-11T21:54:30Z

src/Microsoft.ML.Dnn/DnnUtils.cs

+
+            tf.train.import_meta_graph(path);
+
+            return graph;


remove new lines. #Resolved

eerhardt

Please remove the CMake output files. And ensure this doesn't regress performance with the calls to ToArray.

eerhardt · 2019-09-11T21:44:44Z

docs/samples/Microsoft.ML.Samples/Dynamic/TensorFlow/ImageClassification.cs

@@ -34,11 +34,13 @@ public static void Example()
            var idv = mlContext.Data.LoadFromEnumerable(data);

            // Create a ML pipeline.
+


I'm not sure if these changes are intended. Can they be reverted? #Resolved

eerhardt · 2019-09-11T21:46:00Z

src/Microsoft.ML.Dnn/DnnRetrainTransform.cs

+                var newShape = new int[tfInputShape.ndim];
+                var dims = tfInputShape.dims;
+                newShape[0] = dims[0] == 0 || dims[0] == -1 ? batchSize : dims[0];
+                //var v = tfInputShape[0].Equals(0);


Is this comment useful? #Resolved

eerhardt · 2019-09-11T21:46:29Z

src/Microsoft.ML.Dnn/DnnRetrainTransform.cs

-            loss = tensor.Length > 0 && tensor[0] != IntPtr.Zero ? (float)tensor[0].Data<float>()[0] : 0.0f;
-            metric = tensor.Length > 1 && tensor[1] != IntPtr.Zero ? (float)tensor[1].Data<float>()[0] : 0.0f;
-            var b = tensor.Length > 2 && tensor[2] != IntPtr.Zero ? (float[])tensor[2].Data<float>() : null;
+            //var buffer = tensor[0].Data();


Is this comment useful? #Resolved

src/Microsoft.ML.Dnn/DnnRetrainTransform.cs

eerhardt · 2019-09-11T21:48:55Z

src/Microsoft.ML.Dnn/DnnRetrainTransform.cs

+                        var l = new int[originalShape.ndim];
+                        for (int ishape = 0; ishape < originalShape.ndim; ishape++)
+                        {
+                            var dims = originalShape.dims;


I think this line can be lifted outside of the for loop. No reason to do it every time through the loop. #Resolved

eerhardt · 2019-09-11T21:49:39Z

src/Native/out/build/x64-Debug/.cmake/api/v1/query/client-MicrosoftVS/query.json

@@ -0,0 +1 @@
+{"requests":[{"kind":"cache","version":2},{"kind":"cmakeFiles","version":1},{"kind":"codemodel","version":2}]}


I assume all changes in src\Native are not intended. Please revert. #Resolved

src/Microsoft.ML.Dnn/ImageClassificationTransform.cs

eerhardt · 2019-09-11T21:52:12Z

src/Microsoft.ML.Dnn/ImageClassificationTransform.cs

@@ -1023,7 +1018,7 @@ internal sealed class Options : TransformInputBase
            /// <summary>
            /// Location of the TensorFlow model.
            /// </summary>
-            [Argument(ArgumentType.Required, HelpText = "TensorFlow model used by the transform. Please see https://www.tensorflow.org/mobile/prepare_models for more details.", SortOrder = 0)]
+            [Argument(ArgumentType.Required, HelpText = "TensorFlow model used by the transform. Please see https://www.tf.org/mobile/prepare_models for more details.", SortOrder = 0)]


I can't access this new URL. Is it correct? #Resolved

eerhardt · 2019-09-11T21:52:57Z

src/Microsoft.ML.Dnn/Microsoft.ML.Dnn.csproj

@@ -16,7 +16,7 @@
    <PackageReference Include="SciSharp.TensorFlow.Redist" Version="1.14.0" />
    <PackageReference Include="System.IO.FileSystem.AccessControl" Version="$(SystemIOFileSystemAccessControl)" />
    <PackageReference Include="System.Security.Principal.Windows" Version="$(SystemSecurityPrincipalWindows)" />
-    <PackageReference Include="TensorFlow.NET" Version="0.10.10" />
+    <PackageReference Include="TensorFlow.NET" Version="0.11.3" />


Don't forget to also update the file in the pkg. Usually we put all the dependency versions in a single file - https://github.com/dotnet/machinelearning/blob/master/build/Dependencies.props. That way they only need to be updated in a single place. #Resolved

eerhardt · 2019-09-11T21:53:22Z

src/Microsoft.ML.TensorFlow/TensorflowTransform.cs

@@ -428,7 +428,7 @@ private protected override void SaveModel(ModelSaveContext ctx)
                var buffer = Session.graph.ToGraphDef(status);
                ctx.SaveBinaryStream("TFModel", w =>
                {
-                    w.WriteByteArray(buffer.Data);
+                    w.WriteByteArray(buffer.MemoryBlock.ToArray());


Same comment about ToArray

I'll stop commenting on them, and assume you will test the perf before and after these changes. And fix any perf degradations. #Resolved

codemzs · 2019-09-11T21:55:47Z

src/Microsoft.ML.Dnn/ImageClassificationTransform.cs

@@ -530,14 +526,14 @@ private void VariableSummaries(RefVariable var)
        private (Operation, Tensor, Tensor, Tensor) AddFinalRetrainOps(int classCount, string labelColumn,
            string scoreColumnName, float learningRate, Tensor bottleneckTensor, bool isTraining)
        {
-            var (batch_size, bottleneck_tensor_size) = (bottleneckTensor.TensorShape.Dimensions[0], bottleneckTensor.TensorShape.Dimensions[1]);
+            var (batch_size, bottleneck_tensor_size) = (bottleneckTensor.TensorShape.dims[0], bottleneckTensor.TensorShape.dims[1]);


dims [](start = 85, length = 4)

cache #Resolved

codemzs · 2019-09-11T21:58:26Z

src/Microsoft.ML.Dnn/ImageClassificationTransform.cs

@@ -757,7 +752,7 @@ private protected override void SaveModel(ModelSaveContext ctx)
            var buffer = _session.graph.ToGraphDef(status);
            ctx.SaveBinaryStream("TFModel", w =>
            {
-                w.WriteByteArray(buffer.Data);
+                w.WriteByteArray(buffer.MemoryBlock.ToArray());


w.WriteByteArray(buffer.MemoryBlock.ToArray()); [](start = 16, length = 47)

can you please do a diff between the graph with your change and master and see if there are any changes? I know the graph size increased and I want to know why and by how much. #Resolved

The size of the files are exactly the same between the master and my change, but doing a diff shows that the files differ. Are the trained values expected to be exactly the same?

In reply to: 323478331 [](ancestors = 323478331)

codemzs · 2019-09-11T21:59:34Z

src/Microsoft.ML.TensorFlow/TensorflowTransform.cs

@@ -380,7 +380,7 @@ internal static (TF_DataType[] tfOutputTypes, DataViewType[] outputTypes, (Opera
                // i.e. the first dimension (if unknown) is assumed to be batch dimension.
                // If there are other dimension that are unknown the transformer will return a variable length vector.
                // This is the work around in absence of reshape transformer.
-                int[] dims = shape.NDim > 0 ? shape.Dimensions.Skip(shape[0] == -1 ? 1 : 0).ToArray() : new[] { 0 };
+                int[] dims = shape.ndim > 0 ? shape.dims.Skip(shape.dims[0] == -1 ? 1 : 0).ToArray() : new[] { 0 };


dims [](start = 52, length = 4)

cache #Resolved

codemzs · 2019-09-11T22:00:44Z

src/Microsoft.ML.TensorFlow/TensorflowTransform.cs

-                        for (int ishape = 0; ishape < originalShape.NDim; ishape++)
-                            l[ishape] = originalShape[ishape] == -1 ? (int)d : originalShape[ishape];
+                        var l = new int[originalShape.ndim];
+                        for (int ishape = 0; ishape < originalShape.ndim; ishape++)


ndim [](start = 68, length = 4)

cache #Resolved

codemzs · 2019-09-11T22:00:54Z

src/Microsoft.ML.TensorFlow/TensorflowTransform.cs

-                            l[ishape] = originalShape[ishape] == -1 ? (int)d : originalShape[ishape];
+                        var l = new int[originalShape.ndim];
+                        for (int ishape = 0; ishape < originalShape.ndim; ishape++)
+                            l[ishape] = originalShape.dims[ishape] == -1 ? (int)d : originalShape.dims[ishape];


dims [](start = 98, length = 4)

cache #Resolved

codemzs · 2019-09-11T22:02:41Z

src/Microsoft.ML.TensorFlow/TensorflowTransform.cs

@@ -635,6 +635,7 @@ private void UpdateCacheIfNeeded(long position, ITensorValueGetter[] srcTensorGe
            {
                if (outputCache.Position != position)
                {
+                    _parent.Session.graph.as_default();


_parent.Session.graph.as_default(); [](start = 20, length = 35)

This line is setting the graph as the default graph every time the cursor moves to the new row. We just need to do this once in the thread of the cursor. Please fix, this is call is an interop call and can reduce performance at scoring. #Resolved

codemzs

🕐

safern · 2019-09-12T17:05:50Z

docs/samples/Microsoft.ML.Samples/Dynamic/TensorFlow/ImageClassification.cs

@@ -38,7 +38,7 @@ public static void Example()
                .ScoreTensorFlowModel(
                new[] { nameof(OutputScores.output) },
                new[] { nameof(TensorData.input) }, addBatchDimensionInput: true);
-
+             


Nit: tabs in empty line. #Resolved

…ault graph.

eerhardt · 2019-09-13T18:33:05Z

src/Microsoft.ML.Dnn/DnnRetrainTransform.cs

-            var b = tensor.Length > 2 && tensor[2] != IntPtr.Zero ? (float[])tensor[2].Data<float>() : null;
+            loss = tensor.Length > 0 && tensor[0] != IntPtr.Zero ? (float)tensor[0].ToArray<float>()[0] : 0.0f;
+            metric = tensor.Length > 1 && tensor[1] != IntPtr.Zero ? (float)tensor[1].ToArray<float>()[0] : 0.0f;
+            var b = tensor.Length > 2 && tensor[2] != IntPtr.Zero ? (float[])tensor[2].ToArray<float>() : null;


I don't see b being used anywhere. Can it be removed? #Resolved

eerhardt · 2019-09-13T18:33:30Z

docs/samples/Microsoft.ML.Samples/Dynamic/TensorFlow/ImageClassification.cs

@@ -38,7 +38,7 @@ public static void Example()
                .ScoreTensorFlowModel(
                new[] { nameof(OutputScores.output) },
                new[] { nameof(TensorData.input) }, addBatchDimensionInput: true);
-
+            


Can you revert this file? Changing files unnecessarily bloats the history. #Resolved

eerhardt · 2019-09-13T18:34:38Z

src/Microsoft.ML.Dnn/DnnRetrainTransform.cs

@@ -857,7 +861,7 @@ protected override Delegate MakeGetter(DataViewRow input, int iinfo, Func<int, b
                return Utils.MarshalInvoke(MakeGetter<int>, type, input, iinfo, srcTensorGetters, activeOutputColNames, outputCache);
            }

-            private Delegate MakeGetter<T>(DataViewRow input, int iinfo, ITensorValueGetter[] srcTensorGetters, string[] activeOutputColNames, OutputCache outputCache)
+            private Delegate MakeGetter<T>(DataViewRow input, int iinfo, ITensorValueGetter[] srcTensorGetters, string[] activeOutputColNames, OutputCache outputCache) where T: unmanaged


where T: unmanaged

Can this ever be used with a string type? If so, this is going to fail at runtime since we use reflection to call this method. #Resolved

I have made a separate function for when T is string to avoid this failure.

In reply to: 324317321 [](ancestors = 324317321)

eerhardt · 2019-09-13T18:42:19Z

src/Microsoft.ML.TensorFlow/Microsoft.ML.TensorFlow.csproj

@@ -10,7 +10,7 @@
  <ItemGroup>
    <PackageReference Include="System.IO.FileSystem.AccessControl" Version="$(SystemIOFileSystemAccessControl)" />
    <PackageReference Include="System.Security.Principal.Windows" Version="$(SystemSecurityPrincipalWindows)" />
-    <PackageReference Include="TensorFlow.NET" Version="0.10.10" />
+    <PackageReference Include="TensorFlow.NET" Version="0.11.3" />


This line can actually be removed. Since this project references Microsoft.ML.Dnn, and Microsoft.ML.Dnn references TF.NET, we don't need to say that this project references TF.NET. That way one duplicated version number can be removed. #Resolved

the major concerns I had are now addressed

…m Dependencies.props, removed unnecessary dependency

eerhardt · 2019-09-13T20:04:44Z

You still need to update

machinelearning/pkg/Microsoft.ML.Dnn/Microsoft.ML.Dnn.nupkgproj

Line 12 in abfd76d

#Resolved

…T version

yaeldekel · 2019-09-15T07:39:49Z

src/Microsoft.ML.Dnn/DnnRetrainTransform.cs

@@ -18,6 +18,7 @@
 using NumSharp;
 using Tensorflow;
 using static Microsoft.ML.Transforms.Dnn.DnnUtils;


Microsoft.ML.Transforms.Dnn.DnnUtils [](start = 13, length = 36)

This should be ordered alphabetically. #Resolved

VS shows a warning if I reorder. I am guessing the 'using static' derivatives should appear at the end, and they are already ordered alphabetically.

In reply to: 324450008 [](ancestors = 324450008)

yaeldekel · 2019-09-15T07:50:47Z

src/Microsoft.ML.Dnn/DnnRetrainTransform.cs

+                        var originalShapeNdim = originalShape.ndim;
+                        var l = new int[originalShapeNdim];
+                        for (int ishape = 0; ishape < originalShapeNdim; ishape++)
+                            l[ishape] = originalShapeDims[ishape] == -1 ? (int)d : originalShapeDims[ishape];


(int)d [](start = 74, length = 6)

Should this cast also be done outside the loop? #Resolved

This is not an addition in this PR. If required, I can do this change

In reply to: 324450404 [](ancestors = 324450404)

yaeldekel · 2019-09-15T07:53:25Z

src/Microsoft.ML.Dnn/DnnRetrainTransform.cs

@@ -912,6 +915,8 @@ private void UpdateCacheIfNeeded(long position, ITensorValueGetter[] srcTensorGe
            {
                if (outputCache.Position != position)
                {
+                    if (_parent.Graph.graph_key != tf.get_default_graph().graph_key)
+                        _parent._session.graph.as_default();


What do these lines do?
#Resolved

if the current graph is not the tf default graph, this line sets it as the default graph.

In reply to: 324450509 [](ancestors = 324450509)

yaeldekel · 2019-09-15T07:56:37Z

src/Microsoft.ML.Dnn/DnnUtils.cs

-            });
+            var graph = new Graph();
+            graph = graph.as_default();
+            tf.train.import_meta_graph(path);


tf.train.import_meta_graph(path); [](start = 12, length = 33)

It seems like there is no connection between graph and tf, does this line do anything to graph? #Resolved

This creates a new graph and the 'as_default()' sets this as the tf default graph https://github.com/SciSharp/TensorFlow.NET/blob/41432600c8263cf972b13930c31479ffb412fa65/src/TensorFlowNET.Core/Graphs/Graph.cs#L134. The 'graph' is where the meta graph is loaded and returned.

In reply to: 324450631 [](ancestors = 324450631)

yaeldekel · 2019-09-15T07:59:03Z

src/Microsoft.ML.Dnn/ImageClassificationTransform.cs

@@ -170,7 +170,7 @@ private void CheckTrainingParameters(ImageClassificationEstimator.Options option
            Host.CheckNonWhiteSpace(options.LabelColumn, nameof(options.LabelColumn));
            Host.CheckNonWhiteSpace(options.TensorFlowLabel, nameof(options.TensorFlowLabel));

-            if (_session.graph.OperationByName(options.TensorFlowLabel) == null)
+            if (_session.graph.OperationByName(_labelTensor.name.Split(':')[0]) == null)


_labelTensor.name.Split(':')[0] [](start = 47, length = 31)

If this is the name of the operation you are searching for in the graph, then the error message should mention the correct name. In this case, the options argument is not needed here. #Resolved

the options argument is not used anymore. The _labelTensor has the correct name that the error message would need

In reply to: 324450724 [](ancestors = 324450724)

yaeldekel · 2019-09-15T08:01:00Z

src/Microsoft.ML.Dnn/ImageClassificationTransform.cs

-            tf_with(evalGraph.as_default(), graph =>
-            {
-                var (_, _, groundTruthInput, finalTensor) = AddFinalRetrainOps(classCount, options.LabelColumn,
+            evalGraph.as_default();


evalGraph.as_default(); [](start = 12, length = 23)

Does this mutate evalGraph? In DnnUtils.cs line 96 you wrote graph = graph.as_default();. #Resolved

In line 842 in this file there is also a call to as_default() without assignment.

In reply to: 324450796 [](ancestors = 324450796)

This sets the tensorflow default graph to the evalGraph. This does not mutate the evalGraph.

In reply to: 324450796 [](ancestors = 324450796)

…ly be for unmanaged types

codemzs · 2019-09-16T20:58:36Z

src/Microsoft.ML.Dnn/DnnRetrainTransform.cs

@@ -235,7 +236,8 @@ private void CheckTrainingParameters(DnnRetrainEstimator.Options options)
                inputTensor.InputType(index);
            var tfInputShape = ((Tensor)inputTensor).TensorShape;

-            if (isInputVector && (tfInputShape == null || (tfInputShape.NDim == 0)))
+            var numInputDims = tfInputShape.ndim;


var numInputDims = tfInputShape.ndim; [](start = 12, length = 37)

It seems tfInputShape can be null. We should set numInputDims to -1 if tfInputShape is null otherwise assign numInputDims = tfInputShape.ndim; #Closed

codemzs · 2019-09-16T21:19:20Z

src/Microsoft.ML.Dnn/ImageClassificationTransform.cs

@@ -843,6 +839,7 @@ public void UpdateCacheIfNeeded()
            protected override Delegate MakeGetter(DataViewRow input, int iinfo, Func<int, bool> activeOutput, out Action disposer)
            {
                disposer = null;
+                _parent._session.graph.as_default();


_parent._session.graph.as_default(); [](start = 16, length = 36)

not needed. #Resolved

Required for the Microsoft.ML.Scenarios.TensorFlowScenariosTests.TensorFlowImageClassification

In reply to: 324890773 [](ancestors = 324890773)

codemzs · 2019-09-16T21:30:25Z

src/Microsoft.ML.TensorFlow/TensorflowTransform.cs

@@ -580,7 +585,7 @@ protected override Delegate MakeGetter(DataViewRow input, int iinfo, Func<int, b
                return Utils.MarshalInvoke(MakeGetter<int>, type, input, iinfo, srcTensorGetters, activeOutputColNames, outputCache);
            }

-            private Delegate MakeGetter<T>(DataViewRow input, int iinfo, ITensorValueGetter[] srcTensorGetters, string[] activeOutputColNames, OutputCache outputCache)
+            private Delegate MakeGetter<T>(DataViewRow input, int iinfo, ITensorValueGetter[] srcTensorGetters, string[] activeOutputColNames, OutputCache outputCache) where T:unmanaged
            {


refactor code for unmanged as done in ImageClassificationTranform. #Closed

I did the refactor and looks like one of the tests was hitting this code path for string, as the test failed due to an error in the refactor. I corrected it and it passes now, but that implies that the original MakeGetter was not erroring out for the type string even after 'where T: unmanaged' was specified. Do we still want to refactor the code? The test that I am referring to is Microsoft.ML.Scenarios.TensorFlowScenariosTests.TensorFlowStringTest

In reply to: 324894953 [](ancestors = 324894953)

…nariosTests.TensorFlowStringTest test clarified that MakeGetter works for string type

codemzs · 2019-09-17T00:30:22Z

src/Microsoft.ML.Dnn/ImageClassificationTransform.cs

@@ -843,6 +839,7 @@ public void UpdateCacheIfNeeded()
            protected override Delegate MakeGetter(DataViewRow input, int iinfo, Func<int, bool> activeOutput, out Action disposer)
            {
                disposer = null;
+                _parent._session.graph.as_default();


_parent._session.graph.as_default(); [](start = 16, length = 36)

is this needed? #Resolved

Yes, test Microsoft.ML.Scenarios.TensorFlowScenariosTests.TensorFlowImageClassification is failing without this

In reply to: 324938993 [](ancestors = 324938993)

codemzs

ashbhandare added 7 commits September 6, 2019 17:26

updated Tensorflow .Net from 0.10.10 to 0.11, changed decode_jpeg to …

2186651

…decode_image

Changed decode_image to decode_jpeg, modified few API changes for TF.…

df47f1a

…NET0.11.2 to reproduce memory corruption.

Fixed null graph error, reproducer for protobuff size error.

efa7ea8

Final changes with all ML Samples running with TF.NET version 0.11.3

72b5715

removed local file diff.out

9536843

Corrected for running all ML.Samples

cd5be6d

cleanup

3008242

ashbhandare requested a review from a team as a code owner September 11, 2019 21:43

codemzs reviewed Sep 11, 2019

View reviewed changes

bpstark reviewed Sep 11, 2019

View reviewed changes

codemzs reviewed Sep 11, 2019

View reviewed changes

eerhardt previously requested changes Sep 11, 2019

View reviewed changes

codemzs reviewed Sep 11, 2019

View reviewed changes

codemzs requested changes Sep 11, 2019

View reviewed changes

resolved common issurs with PR

579bdb6

safern reviewed Sep 12, 2019

View reviewed changes

ashbhandare added 2 commits September 12, 2019 10:23

removed extra tab

e599ac9

Added a check to set default graph only when current graph is not def…

5eaa5d2

…ault graph.

eerhardt reviewed Sep 13, 2019

View reviewed changes

ashbhandare added 2 commits September 13, 2019 11:51

Changed to that tf.net version dependency can directly be changed fro…

42ec0d4

…m Dependencies.props, removed unnecessary dependency

Removed unused variable, reverted ImageClassification.cs

90f7ffd

added previously missed package dependency variable for Tensorflow.NE…

6f3d33d

…T version

yaeldekel reviewed Sep 15, 2019

View reviewed changes

Created a sepatare function for string type so that MakeGetter can on…

ffc95e3

…ly be for unmanaged types

codemzs reviewed Sep 16, 2019

View reviewed changes

ashbhandare added 2 commits September 16, 2019 15:43

Additional refactoring for MakeGetter for String type

e633904

Reverting the refactoring as the Microsoft.ML.Scenarios.TensorFlowSce…

9c2371c

…nariosTests.TensorFlowStringTest test clarified that MakeGetter works for string type

codemzs reviewed Sep 17, 2019

View reviewed changes

codemzs approved these changes Sep 17, 2019

View reviewed changes

ashbhandare merged commit 862ae84 into dotnet:master Sep 17, 2019

ghost locked as resolved and limited conversation to collaborators Mar 20, 2022

		@@ -34,11 +34,13 @@ public static void Example()
		var idv = mlContext.Data.LoadFromEnumerable(data);

		// Create a ML pipeline.

		@@ -0,0 +1 @@
		{"requests":[{"kind":"cache","version":2},{"kind":"cmakeFiles","version":1},{"kind":"codemodel","version":2}]}

Upgrade TF.Net version from 0.10.10 to 0.11.3 #4205

Upgrade TF.Net version from 0.10.10 to 0.11.3 #4205

Conversation

ashbhandare commented Sep 11, 2019

dnfclas commented Sep 11, 2019 • edited by ashbhandare Loading

codemzs commented Sep 11, 2019 • edited by ashbhandare Loading

codemzs Sep 11, 2019 • edited by ashbhandare Loading

Choose a reason for hiding this comment

codemzs Sep 11, 2019 • edited by ashbhandare Loading

Choose a reason for hiding this comment

codemzs Sep 11, 2019 • edited by ashbhandare Loading

Choose a reason for hiding this comment

codemzs Sep 11, 2019 • edited by ashbhandare Loading

Choose a reason for hiding this comment

bpstark Sep 11, 2019 • edited by ashbhandare Loading

Choose a reason for hiding this comment

bpstark Sep 11, 2019 • edited by ashbhandare Loading

Choose a reason for hiding this comment

codemzs Sep 11, 2019 • edited by ashbhandare Loading

Choose a reason for hiding this comment

eerhardt left a comment

Choose a reason for hiding this comment

eerhardt Sep 11, 2019 • edited by ashbhandare Loading

Choose a reason for hiding this comment

eerhardt Sep 11, 2019 • edited by ashbhandare Loading

Choose a reason for hiding this comment

eerhardt Sep 11, 2019 • edited by ashbhandare Loading

Choose a reason for hiding this comment

eerhardt Sep 11, 2019 • edited by ashbhandare Loading

Choose a reason for hiding this comment

eerhardt Sep 11, 2019 • edited by ashbhandare Loading

Choose a reason for hiding this comment

eerhardt Sep 11, 2019 • edited by ashbhandare Loading

Choose a reason for hiding this comment

eerhardt Sep 11, 2019 • edited by ashbhandare Loading

Choose a reason for hiding this comment

eerhardt Sep 11, 2019 • edited by ashbhandare Loading

Choose a reason for hiding this comment

codemzs Sep 11, 2019 • edited by ashbhandare Loading

Choose a reason for hiding this comment

codemzs Sep 11, 2019 • edited by ashbhandare Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

codemzs Sep 11, 2019 • edited by ashbhandare Loading

Choose a reason for hiding this comment

codemzs Sep 11, 2019 • edited by ashbhandare Loading

Choose a reason for hiding this comment

codemzs Sep 11, 2019 • edited by ashbhandare Loading

Choose a reason for hiding this comment

codemzs Sep 11, 2019 • edited by ashbhandare Loading

Choose a reason for hiding this comment

codemzs left a comment

Choose a reason for hiding this comment

safern Sep 12, 2019 • edited by ashbhandare Loading

Choose a reason for hiding this comment

eerhardt Sep 13, 2019 • edited by ashbhandare Loading

Choose a reason for hiding this comment

eerhardt Sep 13, 2019 • edited by ashbhandare Loading

Choose a reason for hiding this comment

eerhardt Sep 13, 2019 • edited by ashbhandare Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

eerhardt Sep 13, 2019 • edited by ashbhandare Loading

Choose a reason for hiding this comment

eerhardt commented Sep 13, 2019 • edited by ashbhandare Loading

yaeldekel Sep 15, 2019 • edited by ashbhandare Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

yaeldekel Sep 15, 2019 • edited by ashbhandare Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

yaeldekel Sep 15, 2019 • edited by ashbhandare Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

yaeldekel Sep 15, 2019 • edited by ashbhandare Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

yaeldekel Sep 15, 2019 • edited by ashbhandare Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

dnfclas commented Sep 11, 2019 •

edited by ashbhandare

Loading

codemzs commented Sep 11, 2019 •

edited by ashbhandare

Loading

codemzs Sep 11, 2019 •

edited by ashbhandare

Loading

codemzs Sep 11, 2019 •

edited by ashbhandare

Loading

codemzs Sep 11, 2019 •

edited by ashbhandare

Loading

codemzs Sep 11, 2019 •

edited by ashbhandare

Loading

bpstark Sep 11, 2019 •

edited by ashbhandare

Loading

bpstark Sep 11, 2019 •

edited by ashbhandare

Loading

codemzs Sep 11, 2019 •

edited by ashbhandare

Loading

eerhardt Sep 11, 2019 •

edited by ashbhandare

Loading

eerhardt Sep 11, 2019 •

edited by ashbhandare

Loading

eerhardt Sep 11, 2019 •

edited by ashbhandare

Loading

eerhardt Sep 11, 2019 •

edited by ashbhandare

Loading

eerhardt Sep 11, 2019 •

edited by ashbhandare

Loading

eerhardt Sep 11, 2019 •

edited by ashbhandare

Loading

eerhardt Sep 11, 2019 •

edited by ashbhandare

Loading

eerhardt Sep 11, 2019 •

edited by ashbhandare

Loading

codemzs Sep 11, 2019 •

edited by ashbhandare

Loading

codemzs Sep 11, 2019 •

edited by ashbhandare

Loading

codemzs Sep 11, 2019 •

edited by ashbhandare

Loading

codemzs Sep 11, 2019 •

edited by ashbhandare

Loading

codemzs Sep 11, 2019 •

edited by ashbhandare

Loading

codemzs Sep 11, 2019 •

edited by ashbhandare

Loading

safern Sep 12, 2019 •

edited by ashbhandare

Loading

eerhardt Sep 13, 2019 •

edited by ashbhandare

Loading

eerhardt Sep 13, 2019 •

edited by ashbhandare

Loading

eerhardt Sep 13, 2019 •

edited by ashbhandare

Loading

eerhardt Sep 13, 2019 •

edited by ashbhandare

Loading

eerhardt commented Sep 13, 2019 •

edited by ashbhandare

Loading

yaeldekel Sep 15, 2019 •

edited by ashbhandare

Loading

yaeldekel Sep 15, 2019 •

edited by ashbhandare

Loading

yaeldekel Sep 15, 2019 •

edited by ashbhandare

Loading

yaeldekel Sep 15, 2019 •

edited by ashbhandare

Loading

yaeldekel Sep 15, 2019 •

edited by ashbhandare

Loading

yaeldekel Sep 15, 2019 •

edited by ashbhandare

Loading

codemzs Sep 16, 2019 •

edited

Loading

codemzs Sep 16, 2019 •

edited by ashbhandare

Loading

codemzs Sep 16, 2019 •

edited

Loading

ashbhandare Sep 16, 2019 •

edited

Loading

codemzs Sep 17, 2019 •

edited by ashbhandare

Loading