Hello World works.

src/TensorFlowNET.Core/Sessions/BaseSession.cs

@@ -302,7 +302,7 @@ namespace Tensorflow
                         // wired, don't know why we have to start from offset 9.
                         // length in the begin
                         var str = UTF8Encoding.Default.GetString(bytes, 9, bytes[8]);
-                        nd = np.array(str).reshape();
+                        nd = np.array(str);
                         break;
                     case TF_DataType.TF_UINT8:
                         var _bytes = new byte[tensor.size];

src/TensorFlowNET.Core/TensorFlowNET.Core.csproj

@@ -63,7 +63,6 @@ Docs: https://tensorflownet.readthedocs.io</Description>
 
   <ItemGroup>
     <PackageReference Include="Google.Protobuf" Version="3.9.0" />
-    <PackageReference Include="System.Runtime.CompilerServices.Unsafe" Version="4.5.2" />
   </ItemGroup>
 
   <ItemGroup>

src/TensorFlowNET.Core/Tensors/dtypes.cs

@@ -51,12 +51,13 @@ namespace Tensorflow
         }
 
         // "sbyte", "byte", "short", "ushort", "int", "uint", "long", "ulong", "float", "double", "Complex"
-        public static TF_DataType as_dtype(Type type)
+        public static TF_DataType as_dtype(Type type, TF_DataType? dtype = null)
         {
-            TF_DataType dtype = TF_DataType.DtInvalid;
-            
             switch (type.Name)
             {
+                case "Char":
+                    dtype =  dtype ?? TF_DataType.TF_UINT8;
+                    break;
                 case "SByte":
                     dtype = TF_DataType.TF_INT8;
                     break;
@@ -100,7 +101,7 @@ namespace Tensorflow
                     throw new Exception("as_dtype Not Implemented");
             }
 
-            return dtype;
+            return dtype.Value;
         }
 
         public static DataType as_datatype_enum(this TF_DataType type)

src/TensorFlowNET.Core/Tensors/tensor_util.cs

@@ -226,7 +226,7 @@ namespace Tensorflow
                 }
             }
 
-            var numpy_dtype = dtypes.as_dtype(nparray.dtype);
+            var numpy_dtype = dtypes.as_dtype(nparray.dtype, dtype: dtype);
             if (numpy_dtype == TF_DataType.DtInvalid)
                 throw new TypeError($"Unrecognized data type: {nparray.dtype}");
 

src/TensorFlowNET.Core/Tensors/tf.constant.cs

@@ -33,6 +33,14 @@ namespace Tensorflow
                 verify_shape: verify_shape,
                 allow_broadcast: false);
 
+        public static Tensor constant(string value,
+            string name = "Const") => constant_op._constant_impl(value,
+                tf.@string,
+                new int[] { 1 },
+                name,
+                verify_shape: false,
+                allow_broadcast: false);
+
         public static Tensor constant(float value,
             int shape,
             string name = "Const") => constant_op._constant_impl(value,

test/TensorFlowNET.Examples/HelloWorld.cs

@@ -29,8 +29,9 @@ namespace TensorFlowNET.Examples
             {
                 // Run the op
                 var result = sess.run(hello);
-                Console.WriteLine(result.ToString());
-                return result.ToString().Equals(str);
+                string result_string = string.Join("", result.GetData<char>());
+                Console.WriteLine(result_string);
+                return result_string.Equals(str);
             });
         }