tf.data.Dataset.from_tensor #446

Author: Oceania2018

src/TensorFlowNET.Core/Data/DatasetManager.cs

@@ -1,4 +1,5 @@
-using System;
+using NumSharp;
+using System;
 using System.Collections.Generic;
 using System.Text;
 using Tensorflow.Data;
@@ -7,6 +8,20 @@ namespace Tensorflow
 {
     public class DatasetManager
     {
+        public IDatasetV2 from_generator<T>(IEnumerable<T> generator, TF_DataType[] output_types, TensorShape[] output_shapes)
+            => new GeneratorDataset();
+
+        /// <summary>
+        /// Creates a `Dataset` with a single element, comprising the given tensors.
+        /// </summary>
+        /// <param name="tensors"></param>
+        /// <returns></returns>
+        public IDatasetV2 from_tensor(NDArray tensors)
+            => new TensorDataset(tensors);
+
+        public IDatasetV2 from_tensor(Tensor tensors)
+            => new TensorDataset(tensors);
+
         public IDatasetV2 from_tensor_slices(Tensor features, Tensor labels)
             => new TensorSliceDataset(features, labels);
 

src/TensorFlowNET.Core/Data/GeneratorDataset.cs

@@ -0,0 +1,11 @@
+using System;
+using System.Collections.Generic;
+using System.Text;
+
+namespace Tensorflow.Data
+{
+    public class GeneratorDataset : DatasetSource
+    {
+
+    }
+}

src/TensorFlowNET.Core/Data/TensorDataset.cs

@@ -0,0 +1,33 @@
+using NumSharp;
+using System;
+using System.Collections.Generic;
+using System.Linq;
+using System.Text;
+using static Tensorflow.Binding;
+
+namespace Tensorflow
+{
+    /// <summary>
+    /// A `Dataset` with a single element.
+    /// </summary>
+    public class TensorDataset : DatasetSource
+    {
+        public TensorDataset(Tensor element)
+        {
+            _tensors = new[] { element };
+            var batched_spec = _tensors.Select(x => x.ToTensorSpec()).ToArray();
+            structure = batched_spec.Select(x => x._unbatch()).ToArray();
+
+            variant_tensor = ops.tensor_dataset(_tensors, output_shapes);
+        }
+
+        public TensorDataset(NDArray element)
+        {
+            _tensors = new[] { tf.convert_to_tensor(element) };
+            var batched_spec = _tensors.Select(x => x.ToTensorSpec()).ToArray();
+            structure = batched_spec.ToArray();
+
+            variant_tensor = ops.tensor_dataset(_tensors, output_shapes);
+        }
+    }
+}

src/TensorFlowNET.Core/Operations/dataset_ops.cs

@@ -8,6 +8,24 @@ namespace Tensorflow
 {
     public class dataset_ops
     {
+        public Tensor tensor_dataset(Tensor[] components, TensorShape[] output_shapes, string name = null)
+        {
+            if (tf.Context.executing_eagerly())
+            {
+                var results = tf.Runner.TFE_FastPathExecute(tf.Context, tf.Context.DeviceName,
+                    "TensorDataset", name,
+                    null,
+                    new object[]
+                    {
+                        components,
+                        "output_shapes", output_shapes
+                    });
+                return results[0];
+            }
+
+            throw new NotImplementedException("");
+        }
+
         /// <summary>
         /// Creates a dataset that emits each dim-0 slice of `components` once.
         /// </summary>

test/TensorFlowNET.UnitTest/Dataset/DatasetTest.cs

@@ -1,7 +1,9 @@
 using Microsoft.VisualStudio.TestTools.UnitTesting;
 using System;
 using System.Collections.Generic;
+using System.Linq;
 using System.Text;
+using Tensorflow.Keras;
 using Tensorflow.UnitTest;
 using static Tensorflow.Binding;
 
@@ -62,6 +64,21 @@ public void FromTensorSlices()
             Assert.AreEqual(5, n);
         }
 
+        [TestMethod]
+        public void FromTensor()
+        {
+            var X = new[] { 2013, 2014, 2015, 2016, 2017 };
+
+            var dataset = tf.data.Dataset.from_tensor(X);
+            int n = 0;
+            foreach (var x in dataset)
+            {
+                Assert.IsTrue(X.SequenceEqual(x.Item1.ToArray<int>()));
+                n += 1;
+            }
+            Assert.AreEqual(1, n);
+        }
+
         [TestMethod]
         public void Shard()
         {