diff --git a/src/TensorFlowNET.Core/NumPy/NDArray.Implicit.cs b/src/TensorFlowNET.Core/NumPy/NDArray.Implicit.cs
index 664ba7f9..88dca940 100644
--- a/src/TensorFlowNET.Core/NumPy/NDArray.Implicit.cs
+++ b/src/TensorFlowNET.Core/NumPy/NDArray.Implicit.cs
@@ -18,22 +18,22 @@ namespace Tensorflow.NumPy
             => new NDArray(array);
 
         public unsafe static implicit operator bool(NDArray nd)
-            => *(bool*)nd.data;
+            => nd.dtype == TF_DataType.TF_BOOL ? *(bool*)nd.data : NDArrayConverter.Scalar<bool>(nd);
 
         public unsafe static implicit operator byte(NDArray nd)
-            => *(byte*)nd.data;
+            => nd.dtype == TF_DataType.TF_INT8 ? *(byte*)nd.data : NDArrayConverter.Scalar<byte>(nd);
 
         public unsafe static implicit operator int(NDArray nd)
-            => *(int*)nd.data;
+            => nd.dtype == TF_DataType.TF_INT32 ? *(int*)nd.data : NDArrayConverter.Scalar<int>(nd);
 
         public unsafe static implicit operator long(NDArray nd)
-            => *(long*)nd.data;
+            => nd.dtype == TF_DataType.TF_INT64 ? *(long*)nd.data : NDArrayConverter.Scalar<long>(nd);
 
         public unsafe static implicit operator float(NDArray nd)
-            => *(float*)nd.data;
+            => nd.dtype == TF_DataType.TF_FLOAT ? *(float*)nd.data : NDArrayConverter.Scalar<float>(nd);
 
         public unsafe static implicit operator double(NDArray nd)
-            => *(double*)nd.data;
+            => nd.dtype == TF_DataType.TF_DOUBLE ? *(double*)nd.data : NDArrayConverter.Scalar<double>(nd);
 
         public static implicit operator NDArray(bool value)
             => new NDArray(value);
diff --git a/src/TensorFlowNET.Core/NumPy/NDArray.Index.cs b/src/TensorFlowNET.Core/NumPy/NDArray.Index.cs
index 9797e861..28471854 100644
--- a/src/TensorFlowNET.Core/NumPy/NDArray.Index.cs
+++ b/src/TensorFlowNET.Core/NumPy/NDArray.Index.cs
@@ -175,8 +175,8 @@ namespace Tensorflow.NumPy
         unsafe void SetData(NDArray src, Slice[] slices, int[] indices, int currentNDim)
         {
             if (dtype != src.dtype)
-                src = src.astype(dtype);
-                // throw new ArrayTypeMismatchException($"Required dtype {dtype} but {array.dtype} is assigned.");
+                // src = src.astype(dtype);
+                throw new ArrayTypeMismatchException($"Required dtype {dtype} but {src.dtype} is assigned.");
 
             if (!slices.Any())
                 return;
diff --git a/src/TensorFlowNET.Core/NumPy/NDArrayConverter.cs b/src/TensorFlowNET.Core/NumPy/NDArrayConverter.cs
new file mode 100644
index 00000000..6e1b8da1
--- /dev/null
+++ b/src/TensorFlowNET.Core/NumPy/NDArrayConverter.cs
@@ -0,0 +1,34 @@
+using System;
+using System.Collections.Generic;
+using System.Linq;
+using System.Text;
+
+namespace Tensorflow.NumPy
+{
+    public class NDArrayConverter
+    {
+        public unsafe static T Scalar<T>(NDArray nd) where T : unmanaged
+            => nd.dtype switch
+            {
+                TF_DataType.TF_FLOAT => Scalar<T>(*(float*)nd.data),
+                TF_DataType.TF_INT64 => Scalar<T>(*(long*)nd.data),
+                _ => throw new NotImplementedException("")
+            };
+
+        static T Scalar<T>(float input)
+            => Type.GetTypeCode(typeof(T)) switch
+            {
+                TypeCode.Int32 => (T)Convert.ChangeType(input, TypeCode.Int32),
+                TypeCode.Single => (T)Convert.ChangeType(input, TypeCode.Single),
+                _ => throw new NotImplementedException("")
+            };
+
+        static T Scalar<T>(long input)
+            => Type.GetTypeCode(typeof(T)) switch
+            {
+                TypeCode.Int32 => (T)Convert.ChangeType(input, TypeCode.Int32),
+                TypeCode.Single => (T)Convert.ChangeType(input, TypeCode.Single),
+                _ => throw new NotImplementedException("")
+            };
+    }
+}
diff --git a/src/TensorFlowNET.Keras/Engine/DataAdapters/DataHandler.cs b/src/TensorFlowNET.Keras/Engine/DataAdapters/DataHandler.cs
index c9c44b8f..1ddddd11 100644
--- a/src/TensorFlowNET.Keras/Engine/DataAdapters/DataHandler.cs
+++ b/src/TensorFlowNET.Keras/Engine/DataAdapters/DataHandler.cs
@@ -78,7 +78,7 @@ namespace Tensorflow.Keras.Engine.DataAdapters
             _insufficient_data = false;
         }
 
-        int _infer_steps(int steps_per_epoch, IDatasetV2 dataset)
+        long _infer_steps(int steps_per_epoch, IDatasetV2 dataset)
         {
             if (steps_per_epoch > -1)
                 return steps_per_epoch;