TensorHandleDevices

TensorFlow.NET.sln

@@ -13,8 +13,6 @@ Project("{9A19103F-16F7-4668-BE54-9A1E7A4F7556}") = "Tensorflow.Keras", "src\Ten
 EndProject
 Project("{9A19103F-16F7-4668-BE54-9A1E7A4F7556}") = "Tensorflow.Keras.UnitTest", "test\Tensorflow.Keras.UnitTest\Tensorflow.Keras.UnitTest.csproj", "{EB92DD90-6346-41FB-B967-2B33A860AD98}"
 EndProject
-Project("{9A19103F-16F7-4668-BE54-9A1E7A4F7556}") = "Tensorflow.Hub", "src\TensorFlowNET.Hub\Tensorflow.Hub.csproj", "{95B077C1-E21B-486F-8BDD-1C902FE687AB}"
-EndProject
 Global
 	GlobalSection(SolutionConfigurationPlatforms) = preSolution
 		Debug|Any CPU = Debug|Any CPU
@@ -107,22 +105,6 @@ Global
 		{EB92DD90-6346-41FB-B967-2B33A860AD98}.Release|Any CPU.Build.0 = Release|Any CPU
 		{EB92DD90-6346-41FB-B967-2B33A860AD98}.Release|x64.ActiveCfg = Release|Any CPU
 		{EB92DD90-6346-41FB-B967-2B33A860AD98}.Release|x64.Build.0 = Release|Any CPU
-		{95B077C1-E21B-486F-8BDD-1C902FE687AB}.Debug|Any CPU.ActiveCfg = Debug|Any CPU
-		{95B077C1-E21B-486F-8BDD-1C902FE687AB}.Debug|Any CPU.Build.0 = Debug|Any CPU
-		{95B077C1-E21B-486F-8BDD-1C902FE687AB}.Debug|x64.ActiveCfg = Debug|Any CPU
-		{95B077C1-E21B-486F-8BDD-1C902FE687AB}.Debug|x64.Build.0 = Debug|Any CPU
-		{95B077C1-E21B-486F-8BDD-1C902FE687AB}.Debug-Minimal|Any CPU.ActiveCfg = Debug|Any CPU
-		{95B077C1-E21B-486F-8BDD-1C902FE687AB}.Debug-Minimal|Any CPU.Build.0 = Debug|Any CPU
-		{95B077C1-E21B-486F-8BDD-1C902FE687AB}.Debug-Minimal|x64.ActiveCfg = Debug|Any CPU
-		{95B077C1-E21B-486F-8BDD-1C902FE687AB}.Debug-Minimal|x64.Build.0 = Debug|Any CPU
-		{95B077C1-E21B-486F-8BDD-1C902FE687AB}.Publish|Any CPU.ActiveCfg = Debug|Any CPU
-		{95B077C1-E21B-486F-8BDD-1C902FE687AB}.Publish|Any CPU.Build.0 = Debug|Any CPU
-		{95B077C1-E21B-486F-8BDD-1C902FE687AB}.Publish|x64.ActiveCfg = Debug|Any CPU
-		{95B077C1-E21B-486F-8BDD-1C902FE687AB}.Publish|x64.Build.0 = Debug|Any CPU
-		{95B077C1-E21B-486F-8BDD-1C902FE687AB}.Release|Any CPU.ActiveCfg = Release|Any CPU
-		{95B077C1-E21B-486F-8BDD-1C902FE687AB}.Release|Any CPU.Build.0 = Release|Any CPU
-		{95B077C1-E21B-486F-8BDD-1C902FE687AB}.Release|x64.ActiveCfg = Release|Any CPU
-		{95B077C1-E21B-486F-8BDD-1C902FE687AB}.Release|x64.Build.0 = Release|Any CPU
 	EndGlobalSection
 	GlobalSection(SolutionProperties) = preSolution
 		HideSolutionNode = FALSE

src/TensorFlowNET.Core/APIs/tf.math.cs

@@ -261,7 +261,7 @@ namespace Tensorflow
 
 
         public Tensor divide(Tensor a, Tensor b)
-            => gen_math_ops.real_div(a, b);
+            => a / b;
 
         public Tensor sqrt(Tensor a, string name = null) 
             => gen_math_ops.sqrt(a, name);

src/TensorFlowNET.Core/Binding.Util.cs

@@ -77,6 +77,19 @@ namespace Tensorflow
             Console.WriteLine(_tostring(obj));
         }
 
+        public static void print(string format, params object[] objects)
+        {
+            if (!format.Contains("{}"))
+                Console.WriteLine(format, string.Join(" ", objects.Select(x => x.ToString())));
+
+            foreach(var obj in objects)
+            {
+
+            }
+
+            Console.WriteLine(format);
+        }
+
         public static int len(object a)
         {
             switch (a)

src/TensorFlowNET.Core/Device/c_api.device.cs

@@ -37,6 +37,16 @@ namespace Tensorflow
         [DllImport(TensorFlowLibName)]
         public static extern int TF_DeviceListCount(IntPtr list);
 
+        /// <summary>
+        /// Retrieves the type of the device at the given index.
+        /// </summary>
+        /// <param name="list">TF_DeviceList*</param>
+        /// <param name="index">int</param>
+        /// <param name="status">TF_Status*</param>
+        /// <returns></returns>
+        [DllImport(TensorFlowLibName)]
+        public static extern IntPtr TF_DeviceListType(IntPtr list, int index, IntPtr status);
+
         /// <summary>
         /// Deallocates the device list.
         /// </summary>
@@ -44,6 +54,18 @@ namespace Tensorflow
         [DllImport(TensorFlowLibName)]
         public static extern void TF_DeleteDeviceList(IntPtr list);
 
+        /// <summary>
+        /// Create a new TFE_TensorHandle with the same contents as 'h' but placed
+        /// in the memory of the device name 'device_name'.
+        /// </summary>
+        /// <param name="h">TFE_TensorHandle*</param>
+        /// <param name="ctx">TFE_Context*</param>
+        /// <param name="device_name">char*</param>
+        /// <param name="status">TF_Status*</param>
+        /// <returns>TFE_TensorHandle*</returns>
+        [DllImport(TensorFlowLibName)]
+        public static extern IntPtr TFE_TensorHandleCopyToDevice(IntPtr h, IntPtr ctx, string device_name, IntPtr status);
+
         /// <summary>
         /// Retrieves the full name of the device (e.g. /job:worker/replica:0/...)
         /// The return value will be a pointer to a null terminated string. The caller
@@ -54,6 +76,6 @@ namespace Tensorflow
         /// <param name="index"></param>
         /// <param name="status">TF_Status*</param>
         [DllImport(TensorFlowLibName)]
-        public static extern string TF_DeviceListName(IntPtr list, int index, IntPtr status);
+        public static extern IntPtr TF_DeviceListName(IntPtr list, int index, IntPtr status);
     }
 }

src/TensorFlowNET.Core/Eager/c_api.eager.cs

@@ -159,6 +159,28 @@ namespace Tensorflow
         [DllImport(TensorFlowLibName)]
         public static extern int TFE_TensorHandleNumDims(IntPtr h, IntPtr status);
 
+        /// <summary>
+        /// Returns the device of the operation that produced `h`. If `h` was produced by
+        /// a copy, returns the destination device of the copy. Note that the returned
+        /// device name is not always the device holding the tensor handle's memory. If
+        /// you want the latter, use TFE_TensorHandleBackingDeviceName. This function
+        /// will block till the operation that produces `h` has completed.
+        /// </summary>
+        /// <param name="h">TFE_TensorHandle*</param>
+        /// <param name="status">TF_Status*</param>
+        /// <returns></returns>
+        [DllImport(TensorFlowLibName)]
+        public static extern IntPtr TFE_TensorHandleDeviceName(IntPtr h, IntPtr status);
+
+        /// <summary>
+        /// Returns the name of the device in whose memory `h` resides.
+        /// </summary>
+        /// <param name="h">TFE_TensorHandle*</param>
+        /// <param name="status">TF_Status*</param>
+        /// <returns></returns>
+        [DllImport(TensorFlowLibName)]
+        public static extern IntPtr TFE_TensorHandleBackingDeviceName(IntPtr h, IntPtr status);
+
         /// <summary>
         /// 
         /// </summary>

src/TensorFlowNET.Core/Eager/pywrap_tfe_src.cs

@@ -14,7 +14,7 @@ namespace Tensorflow.Eager
             string device_name, 
             string opName, 
             string name, 
-            params Tensor[] inputs)
+            params object[] inputs)
         {
             IntPtr op = IntPtr.Zero;
             var attr_list_sizes = new Dictionary<string, int>();
@@ -42,7 +42,14 @@ namespace Tensorflow.Eager
                     else
                     {
                         // The item is a single item.
-                        AddInputToOp(inputs[i], true, input_arg, op, status);
+                        switch (inputs[i])
+                        {
+                            case Tensor inputTensor:
+                                AddInputToOp(inputTensor, true, input_arg, op, status);
+                                break;
+                            default:
+                                throw new NotImplementedException("");
+                        }
                     }
                 }
 

src/TensorFlowNET.Core/Operations/gen_math_ops.cs

@@ -140,19 +140,14 @@ namespace Tensorflow
             return _op.outputs[0];
         }
 
-        public static EagerTensor add(Tensor x, Tensor y, string name = null)
+        public static Tensor add<Tx, Ty>(Tx x, Ty y, string name = null)
         {
             if (tf.context.executing_eagerly())
             {
-                var _result = pywrap_tfe_src.TFE_Py_FastPathExecute(tf.context, "", "Add", name, new[] { x, y });
+                var _result = pywrap_tfe_src.TFE_Py_FastPathExecute(tf.context, "", "Add", name, x, y);
                 return _result;
             }
 
-            return null;
-        }
-
-        public static Tensor add<Tx, Ty>(Tx x, Ty y, string name = null)
-        {
             var _op = _op_def_lib._apply_op_helper("Add", name, args: new { x, y });
 
             return _op.output;
@@ -469,6 +464,12 @@ namespace Tensorflow
 
         public static Tensor cast(Tensor x, TF_DataType DstT, bool Truncate= false, string name= null)
         {
+            if (tf.context.executing_eagerly())
+            {
+                var _result = pywrap_tfe_src.TFE_Py_FastPathExecute(tf.context, "", "Cast", name, x, "DstT", DstT, "Truncate", Truncate);
+                return _result;
+            }
+
             var _op = _op_def_lib._apply_op_helper("Cast", name, args: new { x, DstT, Truncate });
 
             return _op.outputs[0];
@@ -490,6 +491,12 @@ namespace Tensorflow
 
         public static Tensor sub<Tx, Ty>(Tx x, Ty y, string name = null)
         {
+            if (tf.context.executing_eagerly())
+            {
+                var _result = pywrap_tfe_src.TFE_Py_FastPathExecute(tf.context, "", "Sub", name, x, y);
+                return _result;
+            }
+
             var _op = _op_def_lib._apply_op_helper("Sub", name, args: new { x, y });
 
             return _op.outputs[0];
@@ -535,6 +542,12 @@ namespace Tensorflow
 
         public static Tensor mul<Tx, Ty>(Tx x, Ty y, string name = null)
         {
+            if (tf.context.executing_eagerly())
+            {
+                var _result = pywrap_tfe_src.TFE_Py_FastPathExecute(tf.context, "", "Mul", name, x, y);
+                return _result;
+            }
+
             var _op = _op_def_lib._apply_op_helper("Mul", name, args: new { x, y });
 
             return _op.outputs[0];
@@ -549,6 +562,12 @@ namespace Tensorflow
 
         public static Tensor real_div(Tensor x, Tensor y, string name = null)
         {
+            if (tf.context.executing_eagerly())
+            {
+                var _result = pywrap_tfe_src.TFE_Py_FastPathExecute(tf.context, "", "RealDiv", name, x, y);
+                return _result;
+            }
+
             var _op = _op_def_lib._apply_op_helper("RealDiv", name, args: new { x, y });
 
             return _op.outputs[0];
@@ -570,6 +589,12 @@ namespace Tensorflow
 
         public static Tensor floor_div(Tensor x, Tensor y, string name = null)
         {
+            if (tf.context.executing_eagerly())
+            {
+                var _result = pywrap_tfe_src.TFE_Py_FastPathExecute(tf.context, "", "FloorDiv", name, x, y);
+                return _result;
+            }
+
             var _op = _op_def_lib._apply_op_helper("FloorDiv", name, args: new { x, y });
 
             return _op.outputs[0];

src/TensorFlowNET.Core/Operations/math_ops.cs

@@ -712,6 +712,22 @@ namespace Tensorflow
                 var x_dtype = x.dtype.as_base_dtype();
                 var y_dtype = y.dtype.as_base_dtype();
 
+                if (x_dtype != y_dtype)
+                    throw new TypeError($"x and y must have the same dtype, got {x_dtype} != {y_dtype}");
+
+                var dtype = x_dtype switch
+                {
+                    TF_DataType.TF_UINT8 => TF_DataType.TF_FLOAT,
+                    TF_DataType.TF_INT8 => TF_DataType.TF_FLOAT,
+                    TF_DataType.TF_INT16 => TF_DataType.TF_FLOAT,
+                    TF_DataType.TF_UINT16 => TF_DataType.TF_FLOAT,
+                    TF_DataType.TF_INT32 => TF_DataType.TF_DOUBLE,
+                    TF_DataType.TF_INT64 => TF_DataType.TF_DOUBLE,
+                    _ => x_dtype
+                };
+                x = cast(x, dtype);
+                y = cast(y, dtype);
+
                 return gen_math_ops.real_div(x, y, name: name);
             });
         }

src/TensorFlowNET.Core/Tensors/Tensor.Operators.cs

@@ -130,7 +130,7 @@ namespace Tensorflow
         public static Tensor operator *(double lhs, Tensor rhs) => BinaryOpWrapper("mul", lhs, rhs);
         public static Tensor operator *(Tensor lhs, Complex rhs) => BinaryOpWrapper("mul", lhs, rhs);
         public static Tensor operator *(Complex lhs, Tensor rhs) => BinaryOpWrapper("mul", lhs, rhs);
-        public static Tensor operator /(Tensor lhs, Tensor rhs) => BinaryOpWrapper("div", lhs, rhs);
+        public static Tensor operator /(Tensor lhs, Tensor rhs) => BinaryOpWrapper("truediv", lhs, rhs);
         public static Tensor operator /(Tensor lhs, NDArray rhs) => BinaryOpWrapper("div", lhs, rhs);
         public static Tensor operator /(NDArray lhs, Tensor rhs) => BinaryOpWrapper("div", lhs, rhs);
         public static Tensor operator /(Tensor lhs, sbyte rhs) => BinaryOpWrapper("div", lhs, rhs);

src/TensorFlowNET.Core/ops.cs

@@ -97,6 +97,9 @@ namespace Tensorflow
         /// <returns></returns>
         public static Tensor convert_to_tensor(object value, TF_DataType dtype = TF_DataType.DtInvalid, string name = null, TF_DataType preferred_dtype = TF_DataType.DtInvalid)
         {
+            if (value is Tensor tensor)
+                return tensor;
+
             return convert_to_tensor_v2(value, dtype, preferred_dtype, name);
         }
 

test/TensorFlowNET.UnitTest/CApiTest.cs

@@ -103,6 +103,30 @@ namespace TensorFlowNET.UnitTest
         protected IntPtr TFE_TensorHandleResolve(IntPtr h, IntPtr status)
             => c_api.TFE_TensorHandleResolve(h, status);
 
+        protected string TFE_TensorHandleDeviceName(IntPtr h, IntPtr status)
+            => c_api.StringPiece(c_api.TFE_TensorHandleDeviceName(h, status));
+
+        protected string TFE_TensorHandleBackingDeviceName(IntPtr h, IntPtr status)
+            => c_api.StringPiece(c_api.TFE_TensorHandleBackingDeviceName(h, status));
+
+        protected IntPtr TFE_ContextListDevices(IntPtr ctx, IntPtr status)
+            => c_api.TFE_ContextListDevices(ctx, status);
+
+        protected int TF_DeviceListCount(IntPtr list)
+            => c_api.TF_DeviceListCount(list);
+
+        protected string TF_DeviceListType(IntPtr list, int index, IntPtr status)
+            => c_api.StringPiece(c_api.TF_DeviceListType(list, index, status));
+
+        protected string TF_DeviceListName(IntPtr list, int index, IntPtr status)
+            => c_api.StringPiece(c_api.TF_DeviceListName(list, index, status));
+
+        protected void TF_DeleteDeviceList(IntPtr list)
+            => c_api.TF_DeleteDeviceList(list);
+
+        protected IntPtr TFE_TensorHandleCopyToDevice(IntPtr h, IntPtr ctx, string device_name, IntPtr status)
+            => c_api.TFE_TensorHandleCopyToDevice(h, ctx, device_name, status);
+
         protected unsafe void memcpy(void * src, IntPtr dst, ulong size)
         {
             Buffer.MemoryCopy(src, dst.ToPointer(), size, size);

test/TensorFlowNET.UnitTest/Eager/CApi.Eager.Context.cs

@@ -33,7 +33,7 @@ namespace TensorFlowNET.UnitTest.Eager
                 EXPECT_EQ(TF_OK, TF_GetCode(status), TF_Message(status));
             }
 
-            // c_api.TF_DeleteDeviceList(devices);
+            c_api.TF_DeleteDeviceList(devices);
             c_api.TF_DeleteStatus(status);
         }
     }

test/TensorFlowNET.UnitTest/Eager/CApi.Eager.TensorHandleDevices.cs

@@ -0,0 +1,43 @@
+using Microsoft.VisualStudio.TestTools.UnitTesting;
+using System;
+using Tensorflow;
+using Tensorflow.Eager;
+using Buffer = System.Buffer;
+
+namespace TensorFlowNET.UnitTest.Eager
+{
+    public partial class CApiEagerTest
+    {
+        /// <summary>
+        /// TEST(CAPI, TensorHandleDevices)
+        /// </summary>
+        [TestMethod]
+        public unsafe void TensorHandleDevices()
+        {
+            var status = c_api.TF_NewStatus();
+            var opts = TFE_NewContextOptions();
+            var ctx = TFE_NewContext(opts, status);
+            TFE_DeleteContextOptions(opts);
+            ASSERT_EQ(TF_OK, TF_GetCode(status), TF_Message(status));
+
+            var hcpu = TestMatrixTensorHandle();
+            var device_name = TFE_TensorHandleDeviceName(hcpu, status);
+            ASSERT_EQ(TF_OK, TF_GetCode(status), TF_Message(status));
+            ASSERT_TRUE(device_name.Contains("CPU:0"));
+
+            var backing_device_name = TFE_TensorHandleBackingDeviceName(hcpu, status);
+            ASSERT_EQ(TF_OK, TF_GetCode(status), TF_Message(status));
+            ASSERT_TRUE(backing_device_name.Contains("CPU:0"));
+
+            // Disable the test if no GPU is present.
+            string gpu_device_name = "";
+            if(GetDeviceName(ctx, ref gpu_device_name, "GPU"))
+            {
+                var hgpu = TFE_TensorHandleCopyToDevice(hcpu, ctx, gpu_device_name, status);
+                ASSERT_TRUE(TF_GetCode(status) == TF_OK, TF_Message(status));
+
+                // shape_op = ShapeOp(ctx, hgpu);
+            }
+        }
+    }
+}

test/TensorFlowNET.UnitTest/Eager/CApi.Eager.cs

@@ -42,5 +42,30 @@ namespace TensorFlowNET.UnitTest.Eager
 
             return op;
         }
+
+        bool GetDeviceName(IntPtr ctx, ref string device_name, string device_type)
+        {
+            var status = TF_NewStatus();
+            var devices = TFE_ContextListDevices(ctx, status);
+            CHECK_EQ(TF_OK, TF_GetCode(status), TF_Message(status));
+
+            int num_devices = TF_DeviceListCount(devices);
+            for (int i = 0; i < num_devices; ++i)
+            {
+                var dev_type = TF_DeviceListType(devices, i, status);
+                CHECK_EQ(TF_GetCode(status), TF_OK, TF_Message(status));
+                var dev_name = TF_DeviceListName(devices, i, status);
+                CHECK_EQ(TF_GetCode(status), TF_OK, TF_Message(status));
+                if (dev_type == device_type)
+                {
+                    device_name = dev_name;
+                    TF_DeleteDeviceList(devices);
+                    return true;
+                }
+            }
+
+            TF_DeleteDeviceList(devices);
+            return false;
+        }
     }
 }

test/TensorFlowNET.UnitTest/Tensorflow.UnitTest.csproj

@@ -33,12 +33,11 @@
     <PackageReference Include="Microsoft.NET.Test.Sdk" Version="16.5.0" />
     <PackageReference Include="MSTest.TestAdapter" Version="2.1.0" />
     <PackageReference Include="MSTest.TestFramework" Version="2.1.0" />
-    <PackageReference Include="SciSharp.TensorFlow.Redist" Version="1.15.1" />
+    <PackageReference Include="SciSharp.TensorFlow.Redist-Windows-GPU" Version="1.15.1" />
   </ItemGroup>
 
   <ItemGroup>
     <ProjectReference Include="..\..\src\TensorFlowNET.Core\Tensorflow.Binding.csproj" />
-    <ProjectReference Include="..\..\src\TensorFlowNET.Hub\Tensorflow.Hub.csproj" />
   </ItemGroup>
 
   <ItemGroup>