@@ -68,20 +68,27 @@ public Tensor strided_slice<T>(Tensor input, T[] begin, T[] end, T[] strides = n
         /// <param name="name">A name for the operation (optional)</param>
         /// <returns>if num_or_size_splits is a scalar returns num_or_size_splits Tensor objects;
         /// if num_or_size_splits is a 1-D Tensor returns num_or_size_splits.get_shape[0] Tensor objects resulting from splitting value.</returns>
-        public Tensor[] split(Tensor value, int num_split, Tensor axis, string name = null)
+        public Tensor[] split(Tensor value, int num_split, Axis axis, string name = null)
             => array_ops.split(
                 value: value,
                 num_or_size_splits: num_split,
                 axis: axis,
                 name: name);
 
-        public Tensor[] split(Tensor value, int num_split, int axis, string name = null)
+        public Tensor[] split(Tensor value, int[] num_split, Axis axis, string name = null)
             => array_ops.split(
                 value: value,
                 num_or_size_splits: num_split,
-                axis: ops.convert_to_tensor(axis),
+                axis: axis,
                 name: name);
 
+        //public Tensor[] split(Tensor value, int num_split, Axis axis, string name = null)
+        //    => array_ops.split(
+        //        value: value,
+        //        num_or_size_splits: num_split,
+        //        axis: axis,
+        //        name: name);
+
         public Tensor ensure_shape(Tensor x, Shape shape, string name = null)
         {
             return gen_ops.ensure_shape(x, shape, name);
 
@@ -0,0 +1,39 @@
+using Newtonsoft.Json;
+using System;
+using System.Collections.Generic;
+using System.Text;
+
+namespace Tensorflow.Keras.ArgsDefinition.Rnn
+{
+    public class GRUCellArgs : AutoSerializeLayerArgs
+    {
+        [JsonProperty("units")]
+        public int Units { get; set; }
+        // TODO(Rinne): lack of initialized value of Activation. Merging keras
+        // into tf.net could resolve it.
+        [JsonProperty("activation")]
+        public Activation Activation { get; set; }
+        [JsonProperty("recurrent_activation")]
+        public Activation RecurrentActivation { get; set; }
+        [JsonProperty("use_bias")]
+        public bool UseBias { get; set; } = true;
+        [JsonProperty("dropout")]
+        public float Dropout { get; set; } = .0f;
+        [JsonProperty("recurrent_dropout")]
+        public float RecurrentDropout { get; set; } = .0f;
+        [JsonProperty("kernel_initializer")]
+        public IInitializer KernelInitializer { get; set; }
+        [JsonProperty("recurrent_initializer")]
+        public IInitializer RecurrentInitializer { get; set; }
+        [JsonProperty("bias_initializer")]
+        public IInitializer BiasInitializer { get; set; }
+        [JsonProperty("reset_after")]
+        public bool ResetAfter { get;set; }
+        [JsonProperty("implementation")]
+        public int Implementation { get; set; } = 2;
+
+
+
+    }
+
+}
@@ -246,6 +246,18 @@ public ILayer RNN(
             bool time_major = false
             );
 
+        public IRnnCell GRUCell(
+            int units,
+            string activation = "tanh",
+            string recurrent_activation = "sigmoid",
+            bool use_bias = true,
+            string kernel_initializer = "glorot_uniform",
+            string recurrent_initializer = "orthogonal",
+            string bias_initializer = "zeros",
+            float dropout = 0f,
+            float recurrent_dropout = 0f, 
+            bool reset_after = true);
+            
         public ILayer Subtract();
     }
 }
@@ -873,6 +873,45 @@ public ILayer LSTM(int units,
                     UnitForgetBias = unit_forget_bias
                 });
 
+        /// <summary>
+        /// Cell class for the GRU layer.
+        /// </summary>
+        /// <param name="units"></param>
+        /// <param name="activation"></param>
+        /// <param name="recurrent_activation"></param>
+        /// <param name="use_bias"></param>
+        /// <param name="kernel_initializer"></param>
+        /// <param name="recurrent_initializer"></param>
+        /// <param name="bias_initializer"></param>
+        /// <param name="dropout"></param>
+        /// <param name="recurrent_dropout"></param>
+        /// <param name="reset_after"></param>
+        /// <returns></returns>
+        public IRnnCell GRUCell(
+            int units,
+            string activation = "tanh",
+            string recurrent_activation = "sigmoid",
+            bool use_bias = true,
+            string kernel_initializer = "glorot_uniform",
+            string recurrent_initializer = "orthogonal",
+            string bias_initializer = "zeros",
+            float dropout = 0f,
+            float recurrent_dropout = 0f,
+            bool reset_after = true)
+            => new GRUCell(new GRUCellArgs
+            {
+                Units = units,
+                Activation = keras.activations.GetActivationFromName(activation),
+                RecurrentActivation = keras.activations.GetActivationFromName(recurrent_activation),
+                KernelInitializer = GetInitializerByName(kernel_initializer),
+                RecurrentInitializer = GetInitializerByName(recurrent_initializer),
+                BiasInitializer = GetInitializerByName(bias_initializer),
+                UseBias = use_bias,
+                Dropout = dropout,
+                RecurrentDropout = recurrent_dropout,
+                ResetAfter = reset_after
+            });
+
         /// <summary>
         /// 
         /// </summary>
@@ -983,5 +1022,9 @@ public ILayer Normalization(Shape? input_shape = null, int? axis = -1, float? me
                 Variance = variance,
                 Invert = invert
             });
+
+
+
+
     }
 }