Sequential, Metric base and Metric Util class skeleton methods added

src/TensorFlowNET.Keras/Args.cs

@@ -0,0 +1,29 @@
+using System;
+using System.Collections.Generic;
+using System.Text;
+
+namespace Tensorflow.Keras
+{
+    public class Args
+    {
+        private List<object> args = new List<object>();
+
+        public object this[int index]
+        {
+            get
+            {
+                return args.Count < index ? args[index] : null;
+            }
+        }
+
+        public T Get<T>(int index)
+        {
+            return args.Count < index ? (T)args[index] : default(T);
+        }
+
+        public void Add<T>(T arg)
+        {
+            args.Add(arg);
+        }
+    }
+}

src/TensorFlowNET.Keras/Engine/Node.cs

@@ -4,7 +4,7 @@ using System.Text;
 
 namespace Tensorflow.Keras.Engine
 {
-    class Node
+    public class Node
     {
     }
 }

src/TensorFlowNET.Keras/Engine/Sequential.cs

@@ -4,7 +4,8 @@ using System.Text;
 
 namespace Tensorflow.Keras.Engine
 {
-    class Sequential
+    public class Sequential
     {
+        
     }
 }

src/TensorFlowNET.Keras/Metrics/Metric.cs

@@ -1,10 +1,39 @@
 using System;
+using System.Collections;
 using System.Collections.Generic;
 using System.Text;
 
 namespace Tensorflow.Keras.Metrics
 {
-    public abstract class Metric
+    public abstract class Metric : Layers.Layer
     {
+        public string dtype
+        {
+            get
+            {
+                throw new NotImplementedException();
+            }
+        }
+
+        public Metric(string name, string dtype)
+        {
+            throw new NotImplementedException();
+        }
+
+        public void __new__ (Metric cls, Args args, KwArgs kwargs) => throw new NotImplementedException();
+
+        public Tensor __call__(Metric cls, Args args, KwArgs kwargs) => throw new NotImplementedException();
+
+        public virtual Hashtable get_config() => throw new NotImplementedException();
+
+        public virtual void reset_states() => throw new NotImplementedException();
+
+        public abstract void update_state(Args args, KwArgs kwargs);
+
+        public abstract Tensor result();
+
+        public void add_weight(string name, TensorShape shape= null, VariableAggregation aggregation= VariableAggregation.Sum,
+                                VariableSynchronization synchronization = VariableSynchronization.OnRead, Initializers.Initializer initializer= null, 
+                                string dtype= null) => throw new NotImplementedException();
     }
 }

src/TensorFlowNET.Keras/Models.cs

@@ -1,14 +1,42 @@
-using System;
+using Keras.Layers;
+using System;
+using System.Collections;
 using System.Collections.Generic;
 using System.Text;
+using Tensorflow.Keras.Engine;
 
 namespace Tensorflow.Keras
 {
     class Models
     {
-        public class Model : Keras.Engine.Training.Model
-        {
+        public class Model : Keras.Engine.Training.Model{}
 
-        }
+        public static Layer share_weights(Layer layer) => throw new NotImplementedException();
+
+        private static Layer _clone_layer(Layer layer) => throw new NotImplementedException();
+
+        private static Layer _insert_ancillary_layers(Model model, Layer ancillary_layers, string[] metrics_names, Node[] new_nodes) => throw new NotImplementedException();
+
+        private static Node[] _make_new_nodes(Node[] nodes_by_depth, Func<Layer, Layer> layer_fn, Hashtable layer_map, Hashtable tensor_map) => throw new NotImplementedException();
+
+        private static Model _clone_functional_model(Model model, Tensor[] input_tensors = null, Func<Layer, Layer> layer_fn = null) => throw new NotImplementedException();
+
+        private static (Hashtable, Layer[]) _clone_layers_and_model_config(Model model, Layer[] input_layers, Func<Layer, Layer> layer_fn) => throw new NotImplementedException();
+
+        private static (Layer[], Layer[]) _remove_ancillary_layers(Model model, Hashtable layer_map, Layer[] layers) => throw new NotImplementedException();
+
+        private static Sequential _clone_sequential_model(Model model, Tensor[] input_tensors = null, Func<Layer, Layer> layer_fn = null) => throw new NotImplementedException();
+
+        public static Model clone_model(Model model, Tensor[] input_tensors = null, Func<Layer, Layer> layer_fn = null) => throw new NotImplementedException();
+
+        private static void _in_place_subclassed_model_reset(Model model) => throw new NotImplementedException();
+
+        private static void _reset_build_compile_trackers(Model model) => throw new NotImplementedException();
+
+        public static void in_place_subclassed_model_state_restoration(Model model) => throw new NotImplementedException();
+
+        public static void clone_and_build_model(Model model, Tensor[] input_tensors= null, Tensor[] target_tensors= null, object custom_objects= null,
+                                        bool compile_clone= true, bool in_place_reset= false, VariableV1 optimizer_iterations= null, Hashtable optimizer_config= null)
+            => throw new NotImplementedException();
     }
 }

src/TensorFlowNET.Keras/Ops.cs

@@ -1,10 +0,0 @@
-using System;
-using System.Collections.Generic;
-using System.Text;
-
-namespace Tensorflow.Keras
-{
-    class Ops
-    {
-    }
-}

src/TensorFlowNET.Keras/Utils/MetricsUtils.cs

@@ -1,10 +1,60 @@
 using System;
 using System.Collections.Generic;
+using System.Reflection;
 using System.Text;
 
 namespace Tensorflow.Keras.Utils
 {
-    class MetricsUtils
+    public class MetricsUtils
     {
+        public static class Reduction
+        {
+            public const string SUM = "sum";
+            public const string SUM_OVER_BATCH_SIZE = "sum_over_batch_size";
+            public const string WEIGHTED_MEAN = "weighted_mean";
+        }
+
+        public static class ConfusionMatrix
+        {
+            public const string TRUE_POSITIVES = "tp";
+            public const string FALSE_POSITIVES = "fp";
+            public const string TRUE_NEGATIVES = "tn";
+            public const string FALSE_NEGATIVES = "fn";
+        }
+
+        public static class AUCCurve
+        {
+            public const string ROC = "ROC";
+            public const string PR = "PR";
+
+            public static string from_str(string key) => throw new NotImplementedException();
+        }
+
+        public static class AUCSummationMethod
+        {
+            public const string INTERPOLATION = "interpolation";
+            public const string MAJORING = "majoring";
+            public const string MINORING = "minoring";
+
+            public static string from_str(string key) => throw new NotImplementedException();
+        }
+
+        public static dynamic update_state_wrapper(Func<Args, KwArgs, Func<bool>> update_state_fn) => throw new NotImplementedException();
+
+        public static dynamic result_wrapper(Func<Args, Tensor> result_fn) => throw new NotImplementedException();
+
+        public static WeakReference weakmethod(MethodInfo method) => throw new NotImplementedException();
+
+        public static void assert_thresholds_range(float[] thresholds) => throw new NotImplementedException();
+
+        public static void parse_init_thresholds(float[] thresholds, float default_threshold = 0.5f) => throw new NotImplementedException();
+
+        public static Operation update_confusion_matrix_variables(variables variables_to_update, Tensor y_true, Tensor y_pred, float[] thresholds,
+                                                    int? top_k= null,int? class_id= null, Tensor sample_weight= null, bool multi_label= false,
+                                                    Tensor label_weights= null) => throw new NotImplementedException();
+
+        private static Tensor _filter_top_k(Tensor x, int k) => throw new NotImplementedException();
+
+        private static (Tensor[], Tensor) ragged_assert_compatible_and_get_flat_values(Tensor[] values, Tensor mask = null) => throw new NotImplementedException();
     }
 }