diff --git a/src/TensorFlowNET.Core/Keras/Engine/InputSpec.cs b/src/TensorFlowNET.Core/Keras/Engine/InputSpec.cs
index d84c31e1..c4980e92 100644
--- a/src/TensorFlowNET.Core/Keras/Engine/InputSpec.cs
+++ b/src/TensorFlowNET.Core/Keras/Engine/InputSpec.cs
@@ -9,7 +9,7 @@ namespace Tensorflow.Keras.Engine
     /// </summary>
     public class InputSpec
     {
-        public int ndim;
+        public int? ndim;
         public int? min_ndim;
         Dictionary<int, int> axes;
 
@@ -18,7 +18,7 @@ namespace Tensorflow.Keras.Engine
             int? min_ndim = null,
             Dictionary<int, int> axes = null)
         {
-            this.ndim = ndim.Value;
+            this.ndim = ndim;
             if (axes == null)
                 axes = new Dictionary<int, int>();
             this.axes = axes;
diff --git a/src/TensorFlowNET.Core/Keras/Layers/Dense.cs b/src/TensorFlowNET.Core/Keras/Layers/Dense.cs
index 7bf63413..f27d0963 100644
--- a/src/TensorFlowNET.Core/Keras/Layers/Dense.cs
+++ b/src/TensorFlowNET.Core/Keras/Layers/Dense.cs
@@ -1,5 +1,6 @@
 using System;
 using System.Collections.Generic;
+using System.Linq;
 using System.Text;
 using Tensorflow.Keras.Engine;
 using Tensorflow.Operations.Activation;
@@ -8,11 +9,13 @@ namespace Tensorflow.Keras.Layers
 {
     public class Dense : Tensorflow.Layers.Layer
     {
-        protected int uints;
+        protected int units;
         protected IActivation activation;
         protected bool use_bias;
         protected IInitializer kernel_initializer;
         protected IInitializer bias_initializer;
+        protected RefVariable kernel;
+        protected RefVariable bias;
 
         public Dense(int units,
             IActivation activation,
@@ -21,7 +24,7 @@ namespace Tensorflow.Keras.Layers
             IInitializer kernel_initializer = null,
             IInitializer bias_initializer = null) : base(trainable: trainable)
         {
-            this.uints = units;
+            this.units = units;
             this.activation = activation;
             this.use_bias = use_bias;
             this.kernel_initializer = kernel_initializer;
@@ -29,5 +32,28 @@ namespace Tensorflow.Keras.Layers
             this.supports_masking = true;
             this.input_spec = new InputSpec(min_ndim: 2);
         }
+
+        protected override void build(TensorShape input_shape)
+        {
+            var last_dim = input_shape.Dimensions.Last();
+            var axes = new Dictionary<int, int>();
+            axes[-1] = last_dim;
+            input_spec = new InputSpec(min_ndim: 2, axes: axes);
+            kernel = add_weight(
+                "kernel",
+                shape: new int[] { last_dim, units },
+                initializer: kernel_initializer,
+                dtype: _dtype,
+                trainable: true);
+            if (use_bias)
+                bias = add_weight(
+                  "bias",
+                  shape: new int[] { units },
+                  initializer: bias_initializer,
+                  dtype: _dtype,
+                  trainable: true);
+
+            built = true;
+        }
     }
 }
diff --git a/src/TensorFlowNET.Core/Operations/Initializers/TruncatedNormal.cs b/src/TensorFlowNET.Core/Operations/Initializers/TruncatedNormal.cs
index b4f7197b..ae639c8e 100644
--- a/src/TensorFlowNET.Core/Operations/Initializers/TruncatedNormal.cs
+++ b/src/TensorFlowNET.Core/Operations/Initializers/TruncatedNormal.cs
@@ -24,7 +24,7 @@ namespace Tensorflow.Operations.Initializers
 
         public Tensor call(TensorShape shape, TF_DataType dtype)
         {
-            throw new NotImplementedException("");
+            return random_ops.truncated_normal(shape, mean, stddev, dtype : dtype, seed: seed);
         }
 
         public object get_config()