Added unit test for the change

Harshitha Parnandi Venkata · Harshitha Parnandi Venkata · commit b1e57396bc80 · 2019-09-25T13:24:03.000-07:00
diff --git a/docs/samples/Microsoft.ML.Samples/Dynamic/ImageClassification/ResnetV2101TransferLearningTrainTestSplit.cs b/docs/samples/Microsoft.ML.Samples/Dynamic/ImageClassification/ResnetV2101TransferLearningTrainTestSplit.cs
@@ -28,9 +28,10 @@ public static void Example()
                 "images");
 
             //Download the image set and unzip
-            //string finalImagesFolderName = DownloadImageSet(
-            //    imagesDownloadFolderPath);
-            string finalImagesFolderName = "flower_photos";
+            string finalImagesFolderName = DownloadImageSet(
+                imagesDownloadFolderPath);
+            // Use this for testing on big flower datasets:
+            //string finalImagesFolderName = "flower_photos"; 
             string fullImagesetFolderPath = Path.Combine(
                 imagesDownloadFolderPath, finalImagesFolderName);
 
@@ -192,15 +193,7 @@ public static int LoadDataIntoBuffer(string path, ref VBuffer<Byte> imgData)
 
                 count = (int)fileLength;
                 var editor = VBufferEditor.Create(ref imgData, count);
-                //var buffer = File.ReadAllBytes(path);
-                fs.Read(editor.Values);
-                /*
-                for (int i = 0; i < count; i++)
-                {
-                    //editor.Values[i] = (byte) fs.ReadByte();
-                    editor.Values[i] = buffer[i];
-                }
-                */
+                fs.Read(editor.Values);                
                 imgData = editor.Commit();
 
                 return count;
diff --git a/src/Microsoft.ML.Dnn/ImageClassificationTransform.cs b/src/Microsoft.ML.Dnn/ImageClassificationTransform.cs
@@ -189,7 +189,6 @@ private void CheckTrainingParameters(ImageClassificationEstimator.Options option
             return (jpegData, resizedImage);
         }
 
-
         private static Tensor Encode(VBuffer<byte> buffer, int length)
         {
             var size = c_api.TF_StringEncodedSize((UIntPtr)length);
@@ -251,7 +250,6 @@ private void CacheFeaturizedImagesToDisk(IDataView input, string labelColumnName
                 var imageBufGetter = cursor.GetGetter<VBuffer<byte>>(imageBufColumn);
                 UInt32 label = UInt32.MaxValue;
                 VBuffer<byte> imageBuf = default;
-                ReadOnlyMemory<char> imagePath = default;
                 runner.AddInput(inputTensorName);
                 ImageClassificationMetrics metrics = new ImageClassificationMetrics();
                 metrics.Bottleneck = new BottleneckMetrics();
diff --git a/test/Microsoft.ML.Tests/ScenariosWithDirectInstantiation/TensorflowTests.cs b/test/Microsoft.ML.Tests/ScenariosWithDirectInstantiation/TensorflowTests.cs
@@ -1243,7 +1243,7 @@ public void TensorFlowImageClassification()
             IDataView testDataset = trainTestData.TestSet;
 
             var pipeline = mlContext.Model.ImageClassification(
-                "ImagePath", "Label",
+                "ImageVBuf", "Label",
                 arch: ImageClassificationEstimator.Architecture.ResnetV2101,
                 epoch: 5,
                 batchSize: 5,
@@ -1281,12 +1281,41 @@ public void TensorFlowImageClassification()
             }
         }
 
+        public static int LoadDataIntoBuffer(string path, ref VBuffer<Byte> imgData)
+        {
+            int count = -1;
+            // bufferSize == 1 used to avoid unnecessary buffer in FileStream
+            using (FileStream fs = new FileStream(path, FileMode.Open, FileAccess.Read, FileShare.Read, bufferSize: 1))
+            {
+                long fileLength = fs.Length;
+                if (fileLength > int.MaxValue)
+                    throw new IOException($"File {path} too big to open.");
+                else if (fileLength == 0)
+                {
+                    byte[] _imageBuffer;
+
+                    // Some file systems (e.g. procfs on Linux) return 0 for length even when there's content.
+                    // Thus we need to assume 0 doesn't mean empty.
+                    _imageBuffer = File.ReadAllBytes(path);
+                    count = _imageBuffer.Length;
+                    Console.WriteLine("File length is zero");
+                }
+
+                count = (int)fileLength;
+                var editor = VBufferEditor.Create(ref imgData, count);
+                fs.Read(editor.Values);
+                imgData = editor.Commit();
+
+                return count;
+            }
+        }
+
         public static IEnumerable<ImageData> LoadImagesFromDirectory(string folder,
             bool useFolderNameAsLabel = true)
         {
             var files = Directory.GetFiles(folder, "*",
                 searchOption: SearchOption.AllDirectories);
-
+            VBuffer<Byte> imgData = new VBuffer<byte>();
             foreach (var file in files)
             {
                 if (Path.GetExtension(file) != ".jpg")
@@ -1307,10 +1336,14 @@ public static IEnumerable<ImageData> LoadImagesFromDirectory(string folder,
                     }
                 }
 
+                // Get the buffer of bytes
+                int imgSize = LoadDataIntoBuffer(file, ref imgData);
+
                 yield return new ImageData()
                 {
                     ImagePath = file,
-                    Label = label
+                    Label = label,
+                    ImageVBuf = imgData
                 };
 
             }
@@ -1372,6 +1405,8 @@ public class ImageData
 
             [LoadColumn(1)]
             public string Label;
+
+            public VBuffer<byte> ImageVBuf;
         }
 
         public class ImagePrediction

Original file line number	Diff line number	Diff line change
`@@ -189,7 +189,6 @@ private void CheckTrainingParameters(ImageClassificationEstimator.Options option`
`189`	`189`	`return (jpegData, resizedImage);`
`190`	`190`	`}`
`191`	`191`
`192`		`-`
`193`	`192`	`private static Tensor Encode(VBuffer<byte> buffer, int length)`
`194`	`193`	`{`
`195`	`194`	`var size = c_api.TF_StringEncodedSize((UIntPtr)length);`
`@@ -251,7 +250,6 @@ private void CacheFeaturizedImagesToDisk(IDataView input, string labelColumnName`
`251`	`250`	`var imageBufGetter = cursor.GetGetter<VBuffer<byte>>(imageBufColumn);`
`252`	`251`	`UInt32 label = UInt32.MaxValue;`
`253`	`252`	`VBuffer<byte> imageBuf = default;`
`254`		`- ReadOnlyMemory<char> imagePath = default;`
`255`	`253`	`runner.AddInput(inputTensorName);`
`256`	`254`	`ImageClassificationMetrics metrics = new ImageClassificationMetrics();`
`257`	`255`	`metrics.Bottleneck = new BottleneckMetrics();`