model.get_config for Keras.

src/TensorFlowNET.Console/TensorFlowNET.Console.csproj

@@ -2,7 +2,7 @@
 
   <PropertyGroup>
     <OutputType>Exe</OutputType>
-    <TargetFramework>netcoreapp3.1</TargetFramework>
+    <TargetFramework>net5.0</TargetFramework>
     <RootNamespace>Tensorflow</RootNamespace>
     <AssemblyName>Tensorflow</AssemblyName>
   </PropertyGroup>

src/TensorFlowNET.Core/Keras/Engine/INode.cs

@@ -1,4 +1,6 @@
-using System.Collections.Generic;
+using System;
+using System.Collections.Generic;
+using Tensorflow.Keras.Saving;
 
 namespace Tensorflow.Keras.Engine
 {
@@ -10,5 +12,7 @@ namespace Tensorflow.Keras.Engine
         List<Tensor> KerasInputs { get; set; }
         INode[] ParentNodes { get; }
         IEnumerable<(ILayer, int, int, Tensor)> iterate_inbound();
+        bool is_input { get; }
+        NodeConfig serialize(Func<string, int, string> make_node_key, Dictionary<string, int> node_conversion_map);
     }
 }

src/TensorFlowNET.Core/Keras/Layers/ILayer.cs

@@ -1,4 +1,5 @@
 using System.Collections.Generic;
+using Tensorflow.Keras.ArgsDefinition;
 using Tensorflow.Keras.Engine;
 
 namespace Tensorflow.Keras
@@ -14,5 +15,6 @@ namespace Tensorflow.Keras
         List<IVariableV1> trainable_variables { get; }
         TensorShape output_shape { get; }
         int count_params();
+        LayerArgs get_config();
     }
 }

src/TensorFlowNET.Core/Keras/Saving/LayerConfig.cs

@@ -0,0 +1,16 @@
+using System;
+using System.Collections.Generic;
+using System.Text;
+using Tensorflow.Keras.ArgsDefinition;
+using Tensorflow.Keras.Engine;
+
+namespace Tensorflow.Keras.Saving
+{
+    public class LayerConfig
+    {
+        public string Name { get; set; }
+        public string ClassName { get; set; }
+        public LayerArgs Config { get; set; }
+        public List<INode> InboundNodes { get; set; }
+    }
+}

src/TensorFlowNET.Core/Keras/Saving/ModelConfig.cs

@@ -0,0 +1,15 @@
+using System;
+using System.Collections.Generic;
+using System.Text;
+using Tensorflow.Keras.Engine;
+
+namespace Tensorflow.Keras.Saving
+{
+    public class ModelConfig
+    {
+        public string Name { get; set; }
+        public List<LayerConfig> Layers { get; set; }
+        public List<ILayer> InputLayers { get; set; }
+        public List<ILayer> OutputLayers { get; set; }
+    }
+}

src/TensorFlowNET.Core/Keras/Saving/NodeConfig.cs

@@ -0,0 +1,13 @@
+using System;
+using System.Collections.Generic;
+using System.Text;
+
+namespace Tensorflow.Keras.Saving
+{
+    public class NodeConfig
+    {
+        public string Name { get; set; }
+        public int NodeIndex { get; set; }
+        public int TensorIndex { get; set; }
+    }
+}

src/TensorFlowNET.Core/ModelSaving/ModelSaver.cs

@@ -0,0 +1,25 @@
+using System;
+using System.Collections.Generic;
+using System.Text;
+using Tensorflow.Keras.Engine;
+using Tensorflow.Train;
+
+namespace Tensorflow.ModelSaving
+{
+    public class ModelSaver
+    {
+        public void save(Trackable obj, string export_dir, SaveOptions options = null)
+        {
+            var saved_model = new SavedModel();
+            var meta_graph_def = new MetaGraphDef();
+            saved_model.MetaGraphs.Add(meta_graph_def);
+            _build_meta_graph(obj, export_dir, options, meta_graph_def);
+        }
+
+        void _build_meta_graph(Trackable obj, string export_dir, SaveOptions options,
+            MetaGraphDef meta_graph_def = null)
+        {
+
+        }
+    }
+}

src/TensorFlowNET.Core/ModelSaving/SaveOptions.cs

@@ -0,0 +1,18 @@
+using System;
+using System.Collections.Generic;
+using System.Text;
+
+namespace Tensorflow.ModelSaving
+{
+    /// <summary>
+    /// Options for saving to SavedModel.
+    /// </summary>
+    public class SaveOptions
+    {
+        bool save_debug_info;
+        public SaveOptions(bool save_debug_info = false)
+        {
+            this.save_debug_info = save_debug_info;
+        }
+    }
+}

src/TensorFlowNET.Core/Operations/NnOps/RNNCell.cs

@@ -17,6 +17,7 @@
 using System;
 using System.Collections.Generic;
 using Tensorflow.Keras;
+using Tensorflow.Keras.ArgsDefinition;
 using Tensorflow.Keras.Engine;
 using Tensorflow.Operations;
 using Tensorflow.Util;
@@ -132,5 +133,10 @@ namespace Tensorflow
         {
             throw new NotImplementedException();
         }
+
+        public LayerArgs get_config()
+        {
+            throw new NotImplementedException();
+        }
     }
 }

src/TensorFlowNET.Core/Protobuf/Gen.bat

@@ -27,6 +27,7 @@ protoc -I=%SRC_DIR% --csharp_out=%DST_DIR% tensorflow/core/framework/summary.pro
 protoc -I=%SRC_DIR% --csharp_out=%DST_DIR% tensorflow/core/framework/op_def.proto
 protoc -I=%SRC_DIR% --csharp_out=%DST_DIR% tensorflow/core/protobuf/saver.proto
 protoc -I=%SRC_DIR% --csharp_out=%DST_DIR% tensorflow/core/protobuf/saved_object_graph.proto
+protoc -I=%SRC_DIR% --csharp_out=%DST_DIR% tensorflow/core/protobuf/saved_model.proto
 ECHO Download `any.proto` from https://github.com/protocolbuffers/protobuf/tree/master/src/google/protobuf
 protoc -I=%SRC_DIR% --csharp_out=%DST_DIR% tensorflow/core/protobuf/meta_graph.proto
 protoc -I=%SRC_DIR% --csharp_out=%DST_DIR% tensorflow/core/protobuf/cluster.proto

src/TensorFlowNET.Core/Protobuf/SavedModel.cs

@@ -0,0 +1,210 @@
+// <auto-generated>
+//     Generated by the protocol buffer compiler.  DO NOT EDIT!
+//     source: tensorflow/core/protobuf/saved_model.proto
+// </auto-generated>
+#pragma warning disable 1591, 0612, 3021
+#region Designer generated code
+
+using pb = global::Google.Protobuf;
+using pbc = global::Google.Protobuf.Collections;
+using pbr = global::Google.Protobuf.Reflection;
+using scg = global::System.Collections.Generic;
+namespace Tensorflow {
+
+  /// <summary>Holder for reflection information generated from tensorflow/core/protobuf/saved_model.proto</summary>
+  public static partial class SavedModelReflection {
+
+    #region Descriptor
+    /// <summary>File descriptor for tensorflow/core/protobuf/saved_model.proto</summary>
+    public static pbr::FileDescriptor Descriptor {
+      get { return descriptor; }
+    }
+    private static pbr::FileDescriptor descriptor;
+
+    static SavedModelReflection() {
+      byte[] descriptorData = global::System.Convert.FromBase64String(
+          string.Concat(
+            "Cip0ZW5zb3JmbG93L2NvcmUvcHJvdG9idWYvc2F2ZWRfbW9kZWwucHJvdG8S",
+            "CnRlbnNvcmZsb3caKXRlbnNvcmZsb3cvY29yZS9wcm90b2J1Zi9tZXRhX2dy",
+            "YXBoLnByb3RvIl8KClNhdmVkTW9kZWwSIgoac2F2ZWRfbW9kZWxfc2NoZW1h",
+            "X3ZlcnNpb24YASABKAMSLQoLbWV0YV9ncmFwaHMYAiADKAsyGC50ZW5zb3Jm",
+            "bG93Lk1ldGFHcmFwaERlZkJ7ChhvcmcudGVuc29yZmxvdy5mcmFtZXdvcmtC",
+            "EFNhdmVkTW9kZWxQcm90b3NQAVpIZ2l0aHViLmNvbS90ZW5zb3JmbG93L3Rl",
+            "bnNvcmZsb3cvdGVuc29yZmxvdy9nby9jb3JlL2NvcmVfcHJvdG9zX2dvX3By",
+            "b3Rv+AEBYgZwcm90bzM="));
+      descriptor = pbr::FileDescriptor.FromGeneratedCode(descriptorData,
+          new pbr::FileDescriptor[] { global::Tensorflow.MetaGraphReflection.Descriptor, },
+          new pbr::GeneratedClrTypeInfo(null, null, new pbr::GeneratedClrTypeInfo[] {
+            new pbr::GeneratedClrTypeInfo(typeof(global::Tensorflow.SavedModel), global::Tensorflow.SavedModel.Parser, new[]{ "SavedModelSchemaVersion", "MetaGraphs" }, null, null, null, null)
+          }));
+    }
+    #endregion
+
+  }
+  #region Messages
+  /// <summary>
+  /// SavedModel is the high level serialization format for TensorFlow Models.
+  /// See [todo: doc links, similar to session_bundle] for more information.
+  /// </summary>
+  public sealed partial class SavedModel : pb::IMessage<SavedModel> {
+    private static readonly pb::MessageParser<SavedModel> _parser = new pb::MessageParser<SavedModel>(() => new SavedModel());
+    private pb::UnknownFieldSet _unknownFields;
+    [global::System.Diagnostics.DebuggerNonUserCodeAttribute]
+    public static pb::MessageParser<SavedModel> Parser { get { return _parser; } }
+
+    [global::System.Diagnostics.DebuggerNonUserCodeAttribute]
+    public static pbr::MessageDescriptor Descriptor {
+      get { return global::Tensorflow.SavedModelReflection.Descriptor.MessageTypes[0]; }
+    }
+
+    [global::System.Diagnostics.DebuggerNonUserCodeAttribute]
+    pbr::MessageDescriptor pb::IMessage.Descriptor {
+      get { return Descriptor; }
+    }
+
+    [global::System.Diagnostics.DebuggerNonUserCodeAttribute]
+    public SavedModel() {
+      OnConstruction();
+    }
+
+    partial void OnConstruction();
+
+    [global::System.Diagnostics.DebuggerNonUserCodeAttribute]
+    public SavedModel(SavedModel other) : this() {
+      savedModelSchemaVersion_ = other.savedModelSchemaVersion_;
+      metaGraphs_ = other.metaGraphs_.Clone();
+      _unknownFields = pb::UnknownFieldSet.Clone(other._unknownFields);
+    }
+
+    [global::System.Diagnostics.DebuggerNonUserCodeAttribute]
+    public SavedModel Clone() {
+      return new SavedModel(this);
+    }
+
+    /// <summary>Field number for the "saved_model_schema_version" field.</summary>
+    public const int SavedModelSchemaVersionFieldNumber = 1;
+    private long savedModelSchemaVersion_;
+    /// <summary>
+    /// The schema version of the SavedModel instance. Used for versioning when
+    /// making future changes to the specification/implementation. Initial value
+    /// at release will be 1.
+    /// </summary>
+    [global::System.Diagnostics.DebuggerNonUserCodeAttribute]
+    public long SavedModelSchemaVersion {
+      get { return savedModelSchemaVersion_; }
+      set {
+        savedModelSchemaVersion_ = value;
+      }
+    }
+
+    /// <summary>Field number for the "meta_graphs" field.</summary>
+    public const int MetaGraphsFieldNumber = 2;
+    private static readonly pb::FieldCodec<global::Tensorflow.MetaGraphDef> _repeated_metaGraphs_codec
+        = pb::FieldCodec.ForMessage(18, global::Tensorflow.MetaGraphDef.Parser);
+    private readonly pbc::RepeatedField<global::Tensorflow.MetaGraphDef> metaGraphs_ = new pbc::RepeatedField<global::Tensorflow.MetaGraphDef>();
+    /// <summary>
+    /// One or more MetaGraphs.
+    /// </summary>
+    [global::System.Diagnostics.DebuggerNonUserCodeAttribute]
+    public pbc::RepeatedField<global::Tensorflow.MetaGraphDef> MetaGraphs {
+      get { return metaGraphs_; }
+    }
+
+    [global::System.Diagnostics.DebuggerNonUserCodeAttribute]
+    public override bool Equals(object other) {
+      return Equals(other as SavedModel);
+    }
+
+    [global::System.Diagnostics.DebuggerNonUserCodeAttribute]
+    public bool Equals(SavedModel other) {
+      if (ReferenceEquals(other, null)) {
+        return false;
+      }
+      if (ReferenceEquals(other, this)) {
+        return true;
+      }
+      if (SavedModelSchemaVersion != other.SavedModelSchemaVersion) return false;
+      if(!metaGraphs_.Equals(other.metaGraphs_)) return false;
+      return Equals(_unknownFields, other._unknownFields);
+    }
+
+    [global::System.Diagnostics.DebuggerNonUserCodeAttribute]
+    public override int GetHashCode() {
+      int hash = 1;
+      if (SavedModelSchemaVersion != 0L) hash ^= SavedModelSchemaVersion.GetHashCode();
+      hash ^= metaGraphs_.GetHashCode();
+      if (_unknownFields != null) {
+        hash ^= _unknownFields.GetHashCode();
+      }
+      return hash;
+    }
+
+    [global::System.Diagnostics.DebuggerNonUserCodeAttribute]
+    public override string ToString() {
+      return pb::JsonFormatter.ToDiagnosticString(this);
+    }
+
+    [global::System.Diagnostics.DebuggerNonUserCodeAttribute]
+    public void WriteTo(pb::CodedOutputStream output) {
+      if (SavedModelSchemaVersion != 0L) {
+        output.WriteRawTag(8);
+        output.WriteInt64(SavedModelSchemaVersion);
+      }
+      metaGraphs_.WriteTo(output, _repeated_metaGraphs_codec);
+      if (_unknownFields != null) {
+        _unknownFields.WriteTo(output);
+      }
+    }
+
+    [global::System.Diagnostics.DebuggerNonUserCodeAttribute]
+    public int CalculateSize() {
+      int size = 0;
+      if (SavedModelSchemaVersion != 0L) {
+        size += 1 + pb::CodedOutputStream.ComputeInt64Size(SavedModelSchemaVersion);
+      }
+      size += metaGraphs_.CalculateSize(_repeated_metaGraphs_codec);
+      if (_unknownFields != null) {
+        size += _unknownFields.CalculateSize();
+      }
+      return size;
+    }
+
+    [global::System.Diagnostics.DebuggerNonUserCodeAttribute]
+    public void MergeFrom(SavedModel other) {
+      if (other == null) {
+        return;
+      }
+      if (other.SavedModelSchemaVersion != 0L) {
+        SavedModelSchemaVersion = other.SavedModelSchemaVersion;
+      }
+      metaGraphs_.Add(other.metaGraphs_);
+      _unknownFields = pb::UnknownFieldSet.MergeFrom(_unknownFields, other._unknownFields);
+    }
+
+    [global::System.Diagnostics.DebuggerNonUserCodeAttribute]
+    public void MergeFrom(pb::CodedInputStream input) {
+      uint tag;
+      while ((tag = input.ReadTag()) != 0) {
+        switch(tag) {
+          default:
+            _unknownFields = pb::UnknownFieldSet.MergeFieldFrom(_unknownFields, input);
+            break;
+          case 8: {
+            SavedModelSchemaVersion = input.ReadInt64();
+            break;
+          }
+          case 18: {
+            metaGraphs_.AddEntriesFrom(input, _repeated_metaGraphs_codec);
+            break;
+          }
+        }
+      }
+    }
+
+  }
+
+  #endregion
+
+}
+
+#endregion Designer generated code

src/TensorFlowNET.Keras/Engine/Functional.GetConfig.cs

@@ -0,0 +1,76 @@
+using System;
+using System.Collections.Generic;
+using System.Linq;
+using System.Text;
+using Tensorflow.Keras.Layers;
+using Tensorflow.Keras.Saving;
+using Tensorflow.Keras.Utils;
+using static Tensorflow.Binding;
+
+namespace Tensorflow.Keras.Engine
+{
+    public partial class Functional
+    {
+        public ModelConfig get_config()
+        {
+            return get_network_config();
+        }
+
+        /// <summary>
+        /// Builds the config, which consists of the node graph and serialized layers.
+        /// </summary>
+        ModelConfig get_network_config()
+        {
+            var config = new ModelConfig
+            {
+                Name = name
+            };
+
+            var node_conversion_map = new Dictionary<string, int>();
+            foreach (var layer in _layers)
+            {
+                var kept_nodes = _should_skip_first_node(layer) ? 1 : 0;
+                foreach (var (original_node_index, node) in enumerate(layer.InboundNodes))
+                {
+                    var node_key = _make_node_key(layer.Name, original_node_index);
+                    if (NetworkNodes.Contains(node_key))
+                    {
+                        node_conversion_map[node_key] = kept_nodes;
+                        kept_nodes += 1;
+                    }
+                }
+            }
+
+            var layer_configs = new List<LayerConfig>();
+            foreach (var layer in _layers)
+            {
+                var filtered_inbound_nodes = new List<INode>();
+                foreach (var (original_node_index, node) in enumerate(layer.InboundNodes))
+                {
+                    var node_key = _make_node_key(layer.Name, original_node_index);
+                    if (NetworkNodes.Contains(node_key) && !node.is_input)
+                    {
+                        var node_data = node.serialize(_make_node_key, node_conversion_map);
+                        throw new NotImplementedException("");
+                    }
+                }
+
+                var layer_config = generic_utils.serialize_keras_object(layer);
+                layer_config.Name = layer.Name;
+                layer_config.InboundNodes = filtered_inbound_nodes;
+                layer_configs.Add(layer_config);
+            }
+            config.Layers = layer_configs;
+
+            return config;
+        }
+
+        bool _should_skip_first_node(ILayer layer)
+        {
+            return layer is Functional && layer.Layers[0] is InputLayer;
+        }
+
+        string _make_node_key(string layer_name, int node_index)
+            => $"{layer_name}_ib-{node_index}";
+    }
+}

src/TensorFlowNET.Keras/Engine/Functional.cs

@@ -10,7 +10,7 @@ namespace Tensorflow.Keras.Engine
     /// <summary>
     /// A `Functional` model is a `Model` defined as a directed graph of layers.
     /// </summary>
-    public class Functional : Model
+    public partial class Functional : Model
     {
         TensorShape _build_input_shape;
         bool _compute_output_and_mask_jointly;
@@ -283,7 +283,7 @@ namespace Tensorflow.Keras.Engine
 
             // Propagate to all previous tensors connected to this node.
             nodes_in_progress.Add(node);
-            if (!node.IsInput)
+            if (!node.is_input)
             {
                 foreach (var k_tensor in node.KerasInputs)
                 {
@@ -307,7 +307,7 @@ namespace Tensorflow.Keras.Engine
 
         Tensors run_internal_graph(Tensors inputs, bool training = false, Tensors mask = null)
         {
-            if (mask != null)
+            if (mask == null)
             {
                 Tensor[] masks = new Tensor[inputs.Count()];
                 foreach (var (i, input_t) in enumerate(inputs))
@@ -330,7 +330,7 @@ namespace Tensorflow.Keras.Engine
                 foreach (Node node in nodes)
                 {
                     // Input tensors already exist.
-                    if (node.IsInput)
+                    if (node.is_input)
                         continue;
 
                     var layer_inputs = node.MapArguments(tensor_dict);

src/TensorFlowNET.Keras/Engine/Layer.cs

@@ -19,6 +19,7 @@ using System.Collections.Generic;
 using System.Linq;
 using System.Threading;
 using Tensorflow.Keras.ArgsDefinition;
+using Tensorflow.Keras.Saving;
 using Tensorflow.Keras.Utils;
 using Tensorflow.Train;
 using static Tensorflow.Binding;
@@ -241,5 +242,8 @@ namespace Tensorflow.Keras.Engine
                 return weights;
             }
         }
+
+        public virtual LayerArgs get_config()
+            => throw new NotImplementedException("");
     }
 }

src/TensorFlowNET.Keras/Engine/MetricsContainer.cs

@@ -2,7 +2,7 @@
 using System.Collections.Generic;
 using System.Linq;
 using Tensorflow.Keras.Metrics;
-using static Tensorflow.KerasExt;
+using static Tensorflow.KerasApi;
 
 namespace Tensorflow.Keras.Engine
 {

src/TensorFlowNET.Keras/Engine/Model.Save.cs

@@ -1,13 +1,26 @@
 using System.Collections.Generic;
 using Tensorflow.Keras.Metrics;
+using Tensorflow.ModelSaving;
 
 namespace Tensorflow.Keras.Engine
 {
     public partial class Model
     {
-        public void save(string path)
+        ModelSaver saver = new ModelSaver();
+
+        /// <summary>
+        /// Saves the model to Tensorflow SavedModel or a single HDF5 file.
+        /// </summary>
+        /// <param name="filepath"></param>
+        /// <param name="overwrite"></param>
+        /// <param name="include_optimizer"></param>
+        public void save(string filepath,
+            bool overwrite = true,
+            bool include_optimizer = true,
+            string save_format = "tf",
+            SaveOptions options = null)
         {
-          
+            saver.save(this, filepath);
         }
     }
 }

src/TensorFlowNET.Keras/Engine/Node.Serialize.cs

@@ -0,0 +1,18 @@
+using System;
+using System.Collections.Generic;
+using Tensorflow.Keras.Saving;
+
+namespace Tensorflow.Keras.Engine
+{
+    public partial class Node
+    {
+        /// <summary>
+        /// Serializes `Node` for Functional API's `get_config`.
+        /// </summary>
+        /// <returns></returns>
+        public NodeConfig serialize(Func<string, int, string> make_node_key, Dictionary<string, int> node_conversion_map)
+        {
+            throw new NotImplementedException("");
+        }
+    }
+}

src/TensorFlowNET.Keras/Engine/Node.cs

@@ -41,7 +41,7 @@ namespace Tensorflow.Keras.Engine
         public TensorShape[] output_shapes;
         public List<Tensor> KerasInputs { get; set; } = new List<Tensor>();
         public ILayer Layer { get; set; }
-        public bool IsInput => args.InputTensors == null;
+        public bool is_input => args.InputTensors == null;
         public int[] FlatInputIds { get; set; }
         public int[] FlatOutputIds { get; set; }
         bool _single_positional_tensor_passed => KerasInputs.Count() == 1;

src/TensorFlowNET.Keras/Engine/Sequential.cs

@@ -17,7 +17,7 @@
 using System.Collections.Generic;
 using Tensorflow.Keras.ArgsDefinition;
 using Tensorflow.Keras.Layers;
-using static Tensorflow.KerasExt;
+using static Tensorflow.KerasApi;
 
 namespace Tensorflow.Keras.Engine
 {

src/TensorFlowNET.Keras/KerasApi.cs

@@ -1,84 +1,12 @@
-using System.Collections.Generic;
-using Tensorflow.Keras.ArgsDefinition;
-using Tensorflow.Keras.Datasets;
-using Tensorflow.Keras.Engine;
-using Tensorflow.Keras.Layers;
-using Tensorflow.Keras.Losses;
-using Tensorflow.Keras.Metrics;
-using Tensorflow.Keras.Optimizers;
+using Tensorflow.Keras;
 
-namespace Tensorflow.Keras
+namespace Tensorflow
 {
-    public class KerasApi
+    public static class KerasApi
     {
-        public KerasDataset datasets { get; } = new KerasDataset();
-        public Initializers initializers { get; } = new Initializers();
-        public Regularizers regularizers { get; } = new Regularizers();
-        public LayersApi layers { get; } = new LayersApi();
-        public LossesApi losses { get; } = new LossesApi();
-        public Activations activations { get; } = new Activations();
-        public Preprocessing preprocessing { get; } = new Preprocessing();
-        public BackendImpl backend { get; } = new BackendImpl();
-        public OptimizerApi optimizers { get; } = new OptimizerApi();
-        public MetricsApi metrics { get; } = new MetricsApi();
+        public static KerasInterface Keras(this tensorflow tf)
+            => new KerasInterface();
 
-        public Sequential Sequential(List<ILayer> layers = null,
-                string name = null)
-            => new Sequential(new SequentialArgs
-            {
-                Layers = layers,
-                Name = name
-            });
-
-        /// <summary>
-        /// `Model` groups layers into an object with training and inference features.
-        /// </summary>
-        /// <param name="input"></param>
-        /// <param name="output"></param>
-        /// <returns></returns>
-        public Functional Model(Tensors inputs, Tensors outputs, string name = null)
-            => new Functional(inputs, outputs, name: name);
-
-        /// <summary>
-        /// Instantiate a Keras tensor.
-        /// </summary>
-        /// <param name="shape"></param>
-        /// <param name="batch_size"></param>
-        /// <param name="dtype"></param>
-        /// <param name="name"></param>
-        /// <param name="sparse">
-        /// A boolean specifying whether the placeholder to be created is sparse.
-        /// </param>
-        /// <param name="ragged">
-        /// A boolean specifying whether the placeholder to be created is ragged.
-        /// </param>
-        /// <param name="tensor">
-        /// Optional existing tensor to wrap into the `Input` layer.
-        /// If set, the layer will not create a placeholder tensor.
-        /// </param>
-        /// <returns></returns>
-        public Tensor Input(TensorShape shape = null,
-                int batch_size = -1,
-                TF_DataType dtype = TF_DataType.DtInvalid,
-                string name = null,
-                bool sparse = false,
-                bool ragged = false,
-                Tensor tensor = null)
-        {
-            var args = new InputLayerArgs
-            {
-                Name = name,
-                InputShape = shape,
-                BatchSize = batch_size,
-                DType = dtype,
-                Sparse = sparse,
-                Ragged = ragged,
-                InputTensor = tensor
-            };
-
-            var layer = new InputLayer(args);
-
-            return layer.InboundNodes[0].Outputs;
-        }
+        public static KerasInterface keras { get; } = new KerasInterface();
     }
 }

src/TensorFlowNET.Keras/KerasExtension.cs

@@ -1,12 +0,0 @@
-using Tensorflow.Keras;
-
-namespace Tensorflow
-{
-    public static class KerasExt
-    {
-        public static KerasApi Keras(this tensorflow tf)
-            => new KerasApi();
-
-        public static KerasApi keras { get; } = new KerasApi();
-    }
-}

src/TensorFlowNET.Keras/KerasInterface.cs

@@ -0,0 +1,84 @@
+using System.Collections.Generic;
+using Tensorflow.Keras.ArgsDefinition;
+using Tensorflow.Keras.Datasets;
+using Tensorflow.Keras.Engine;
+using Tensorflow.Keras.Layers;
+using Tensorflow.Keras.Losses;
+using Tensorflow.Keras.Metrics;
+using Tensorflow.Keras.Optimizers;
+
+namespace Tensorflow.Keras
+{
+    public class KerasInterface
+    {
+        public KerasDataset datasets { get; } = new KerasDataset();
+        public Initializers initializers { get; } = new Initializers();
+        public Regularizers regularizers { get; } = new Regularizers();
+        public LayersApi layers { get; } = new LayersApi();
+        public LossesApi losses { get; } = new LossesApi();
+        public Activations activations { get; } = new Activations();
+        public Preprocessing preprocessing { get; } = new Preprocessing();
+        public BackendImpl backend { get; } = new BackendImpl();
+        public OptimizerApi optimizers { get; } = new OptimizerApi();
+        public MetricsApi metrics { get; } = new MetricsApi();
+
+        public Sequential Sequential(List<ILayer> layers = null,
+                string name = null)
+            => new Sequential(new SequentialArgs
+            {
+                Layers = layers,
+                Name = name
+            });
+
+        /// <summary>
+        /// `Model` groups layers into an object with training and inference features.
+        /// </summary>
+        /// <param name="input"></param>
+        /// <param name="output"></param>
+        /// <returns></returns>
+        public Functional Model(Tensors inputs, Tensors outputs, string name = null)
+            => new Functional(inputs, outputs, name: name);
+
+        /// <summary>
+        /// Instantiate a Keras tensor.
+        /// </summary>
+        /// <param name="shape"></param>
+        /// <param name="batch_size"></param>
+        /// <param name="dtype"></param>
+        /// <param name="name"></param>
+        /// <param name="sparse">
+        /// A boolean specifying whether the placeholder to be created is sparse.
+        /// </param>
+        /// <param name="ragged">
+        /// A boolean specifying whether the placeholder to be created is ragged.
+        /// </param>
+        /// <param name="tensor">
+        /// Optional existing tensor to wrap into the `Input` layer.
+        /// If set, the layer will not create a placeholder tensor.
+        /// </param>
+        /// <returns></returns>
+        public Tensor Input(TensorShape shape = null,
+                int batch_size = -1,
+                TF_DataType dtype = TF_DataType.DtInvalid,
+                string name = null,
+                bool sparse = false,
+                bool ragged = false,
+                Tensor tensor = null)
+        {
+            var args = new InputLayerArgs
+            {
+                Name = name,
+                InputShape = shape,
+                BatchSize = batch_size,
+                DType = dtype,
+                Sparse = sparse,
+                Ragged = ragged,
+                InputTensor = tensor
+            };
+
+            var layer = new InputLayer(args);
+
+            return layer.InboundNodes[0].Outputs;
+        }
+    }
+}

src/TensorFlowNET.Keras/Layers/InputLayer.cs

@@ -19,7 +19,7 @@ using Tensorflow.Framework.Models;
 using Tensorflow.Keras.ArgsDefinition;
 using Tensorflow.Keras.Engine;
 using static Tensorflow.Binding;
-using static Tensorflow.KerasExt;
+using static Tensorflow.KerasApi;
 
 namespace Tensorflow.Keras.Layers
 {
@@ -50,6 +50,7 @@ namespace Tensorflow.Keras.Layers
             {
                 var prefix = "input";
                 name = prefix + '_' + keras.backend.get_uid(prefix);
+                args.Name = name;
             }
 
             if (args.DType == TF_DataType.DtInvalid)
@@ -99,5 +100,8 @@ namespace Tensorflow.Keras.Layers
 
             tf.Context.restore_mode();
         }
+
+        public override LayerArgs get_config()
+            => args;
     }
 }

src/TensorFlowNET.Keras/Layers/LayersApi.cs

@@ -2,7 +2,7 @@
 using Tensorflow.Keras.ArgsDefinition;
 using Tensorflow.Keras.Engine;
 using static Tensorflow.Binding;
-using static Tensorflow.KerasExt;
+using static Tensorflow.KerasApi;
 
 namespace Tensorflow.Keras.Layers
 {

src/TensorFlowNET.Keras/Layers/ZeroPadding2D.cs

@@ -2,7 +2,7 @@
 using Tensorflow.Keras.ArgsDefinition;
 using Tensorflow.Keras.Engine;
 using Tensorflow.Keras.Utils;
-using static Tensorflow.KerasExt;
+using static Tensorflow.KerasApi;
 
 namespace Tensorflow.Keras.Layers
 {

src/TensorFlowNET.Keras/Preprocessings/Preprocessing.image_dataset_from_directory.cs

@@ -1,4 +1,4 @@
-using static Tensorflow.KerasExt;
+using static Tensorflow.KerasApi;
 
 namespace Tensorflow.Keras
 {

src/TensorFlowNET.Keras/Tensorflow.Keras.csproj

@@ -35,4 +35,8 @@ Keras is an API designed for human beings, not machines. Keras follows best prac
     <ProjectReference Include="..\TensorFlowNET.Core\Tensorflow.Binding.csproj" />
   </ItemGroup>
 
+  <ItemGroup>
+    <Folder Include="Saving\" />
+  </ItemGroup>
+
 </Project>

src/TensorFlowNET.Keras/Utils/base_layer_utils.cs

@@ -21,7 +21,7 @@ using System.Linq;
 using Tensorflow.Keras.ArgsDefinition;
 using Tensorflow.Keras.Engine;
 using static Tensorflow.Binding;
-using static Tensorflow.KerasExt;
+using static Tensorflow.KerasApi;
 
 namespace Tensorflow.Keras.Utils
 {

src/TensorFlowNET.Keras/Utils/generic_utils.cs

@@ -16,11 +16,22 @@
 
 using System;
 using System.Linq;
+using Tensorflow.Keras.Saving;
 
 namespace Tensorflow.Keras.Utils
 {
     public class generic_utils
     {
+        public static LayerConfig serialize_keras_object(ILayer instance)
+        {
+            var config = instance.get_config();
+            return new LayerConfig
+            {
+                Config = config,
+                ClassName = instance.GetType().Name
+            };
+        }
+
         public static string to_snake_case(string name)
         {
             return string.Concat(name.Select((x, i) =>

src/TensorFlowNet.Benchmarks/Tensorflow.Benchmark.csproj

@@ -2,7 +2,7 @@
 
   <PropertyGroup>
     <OutputType>Exe</OutputType>
-    <TargetFramework>netcoreapp3.1</TargetFramework>
+    <TargetFramework>net5.0</TargetFramework>
     <Platforms>AnyCPU;x64</Platforms>
   </PropertyGroup>
 

test/TensorFlowNET.UnitTest/Keras/LayersTest.cs

@@ -1,6 +1,6 @@
 using Microsoft.VisualStudio.TestTools.UnitTesting;
 using NumSharp;
-using static Tensorflow.KerasExt;
+using static Tensorflow.KerasApi;
 
 namespace TensorFlowNET.UnitTest.Keras
 {

test/TensorFlowNET.UnitTest/Keras/ModelSaveTest.cs

@@ -1,6 +1,6 @@
 using Microsoft.VisualStudio.TestTools.UnitTesting;
 using Tensorflow.Keras.Engine;
-using static Tensorflow.KerasExt;
+using static Tensorflow.KerasApi;
 
 namespace TensorFlowNET.UnitTest.Keras
 {
@@ -10,21 +10,20 @@ namespace TensorFlowNET.UnitTest.Keras
     [TestClass]
     public class ModelSaveTest : EagerModeTestBase
     {
-        [TestMethod]
-        public void SaveAndLoadTest()
+        [TestMethod, Ignore]
+        public void GetAndFromConfig()
         {
-            var model = GetModel();
+            var model = GetFunctionalModel();
+            var config = model.get_config();
         }
 
-        Model GetModel()
+        Functional GetFunctionalModel()
         {
             // Create a simple model.
             var inputs = keras.Input(shape: 32);
             var dense_layer = keras.layers.Dense(1);
             var outputs = dense_layer.Apply(inputs);
-            var model = keras.Model(inputs, outputs);
-            model.compile("adam", "mean_squared_error");
-            return model;
+            return keras.Model(inputs, outputs);
         }
     }
 }