Seperated Keras from binding project.

TensorFlow.NET.sln

@@ -11,6 +11,8 @@ Project("{9A19103F-16F7-4668-BE54-9A1E7A4F7556}") = "Tensorflow.UnitTest", "test
 EndProject
 Project("{9A19103F-16F7-4668-BE54-9A1E7A4F7556}") = "TensorFlowNET.Console", "src\TensorFlowNET.Console\TensorFlowNET.Console.csproj", "{03F06299-3F4B-4449-A709-3A647657BC0C}"
 EndProject
+Project("{9A19103F-16F7-4668-BE54-9A1E7A4F7556}") = "Tensorflow.Keras", "src\TensorFlowNET.Keras\Tensorflow.Keras.csproj", "{49D71826-C03D-4FA7-9BAC-22C1327E65CF}"
+EndProject
 Global
 	GlobalSection(SolutionConfigurationPlatforms) = preSolution
 		Debug|Any CPU = Debug|Any CPU
@@ -123,6 +125,30 @@ Global
 		{03F06299-3F4B-4449-A709-3A647657BC0C}.Release|x64.Build.0 = Release|Any CPU
 		{03F06299-3F4B-4449-A709-3A647657BC0C}.Release|x86.ActiveCfg = Release|Any CPU
 		{03F06299-3F4B-4449-A709-3A647657BC0C}.Release|x86.Build.0 = Release|Any CPU
+		{49D71826-C03D-4FA7-9BAC-22C1327E65CF}.Debug|Any CPU.ActiveCfg = Debug|Any CPU
+		{49D71826-C03D-4FA7-9BAC-22C1327E65CF}.Debug|Any CPU.Build.0 = Debug|Any CPU
+		{49D71826-C03D-4FA7-9BAC-22C1327E65CF}.Debug|x64.ActiveCfg = Debug|x64
+		{49D71826-C03D-4FA7-9BAC-22C1327E65CF}.Debug|x64.Build.0 = Debug|x64
+		{49D71826-C03D-4FA7-9BAC-22C1327E65CF}.Debug|x86.ActiveCfg = Debug|Any CPU
+		{49D71826-C03D-4FA7-9BAC-22C1327E65CF}.Debug|x86.Build.0 = Debug|Any CPU
+		{49D71826-C03D-4FA7-9BAC-22C1327E65CF}.Debug-Minimal|Any CPU.ActiveCfg = Debug|Any CPU
+		{49D71826-C03D-4FA7-9BAC-22C1327E65CF}.Debug-Minimal|Any CPU.Build.0 = Debug|Any CPU
+		{49D71826-C03D-4FA7-9BAC-22C1327E65CF}.Debug-Minimal|x64.ActiveCfg = Debug|x64
+		{49D71826-C03D-4FA7-9BAC-22C1327E65CF}.Debug-Minimal|x64.Build.0 = Debug|x64
+		{49D71826-C03D-4FA7-9BAC-22C1327E65CF}.Debug-Minimal|x86.ActiveCfg = Debug|Any CPU
+		{49D71826-C03D-4FA7-9BAC-22C1327E65CF}.Debug-Minimal|x86.Build.0 = Debug|Any CPU
+		{49D71826-C03D-4FA7-9BAC-22C1327E65CF}.Publish|Any CPU.ActiveCfg = Release|Any CPU
+		{49D71826-C03D-4FA7-9BAC-22C1327E65CF}.Publish|Any CPU.Build.0 = Release|Any CPU
+		{49D71826-C03D-4FA7-9BAC-22C1327E65CF}.Publish|x64.ActiveCfg = Debug|x64
+		{49D71826-C03D-4FA7-9BAC-22C1327E65CF}.Publish|x64.Build.0 = Debug|x64
+		{49D71826-C03D-4FA7-9BAC-22C1327E65CF}.Publish|x86.ActiveCfg = Release|Any CPU
+		{49D71826-C03D-4FA7-9BAC-22C1327E65CF}.Publish|x86.Build.0 = Release|Any CPU
+		{49D71826-C03D-4FA7-9BAC-22C1327E65CF}.Release|Any CPU.ActiveCfg = Release|Any CPU
+		{49D71826-C03D-4FA7-9BAC-22C1327E65CF}.Release|Any CPU.Build.0 = Release|Any CPU
+		{49D71826-C03D-4FA7-9BAC-22C1327E65CF}.Release|x64.ActiveCfg = Release|x64
+		{49D71826-C03D-4FA7-9BAC-22C1327E65CF}.Release|x64.Build.0 = Release|x64
+		{49D71826-C03D-4FA7-9BAC-22C1327E65CF}.Release|x86.ActiveCfg = Release|Any CPU
+		{49D71826-C03D-4FA7-9BAC-22C1327E65CF}.Release|x86.Build.0 = Release|Any CPU
 	EndGlobalSection
 	GlobalSection(SolutionProperties) = preSolution
 		HideSolutionNode = FALSE

src/TensorFlowNET.Console/Program.cs

@@ -1,5 +1,4 @@
 using System;
-using static Tensorflow.Binding;
 
 namespace Tensorflow
 {

src/TensorFlowNET.Console/TensorFlowNET.Console.csproj

@@ -12,7 +12,7 @@
   </ItemGroup>
 
   <ItemGroup>
-    <ProjectReference Include="..\TensorFlowNET.Core\Tensorflow.Binding.csproj" />
+    <ProjectReference Include="..\TensorFlowNET.Keras\Tensorflow.Keras.csproj" />
   </ItemGroup>
 
 </Project>

src/TensorFlowNET.Core/APIs/tf.nn.cs

@@ -137,8 +137,6 @@ namespace Tensorflow
                     is_training: is_training,
                     name: name);
 
-            public IPoolFunction max_pool_fn => new MaxPoolFunction();
-
             public Tensor max_pool(Tensor value, int[] ksize, int[] strides, string padding, string data_format = "NHWC", string name = null)
                 => nn_ops.max_pool(value, ksize, strides, padding, data_format: data_format, name: name);
 

src/TensorFlowNET.Core/APIs/tf.optimizers.cs

@@ -1,42 +0,0 @@
-/*****************************************************************************
-   Copyright 2018 The TensorFlow.NET Authors. All Rights Reserved.
-
-   Licensed under the Apache License, Version 2.0 (the "License");
-   you may not use this file except in compliance with the License.
-   You may obtain a copy of the License at
-
-       http://www.apache.org/licenses/LICENSE-2.0
-
-   Unless required by applicable law or agreed to in writing, software
-   distributed under the License is distributed on an "AS IS" BASIS,
-   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-   See the License for the specific language governing permissions and
-   limitations under the License.
-******************************************************************************/
-
-using Tensorflow.Keras.Optimizers;
-
-namespace Tensorflow
-{
-    public partial class tensorflow
-    {
-        public KerasOptimizers optimizers => new KerasOptimizers();
-
-        public class KerasOptimizers
-        {
-            public SGD SGD(float learning_rate) => new SGD(learning_rate);
-
-            public Adam Adam(float learning_rate = 0.001f,
-                float beta_1 = 0.9f,
-                float beta_2 = 0.999f,
-                float epsilon = 1e-7f,
-                bool amsgrad = false,
-                string name = "Adam") => new Adam(learning_rate: learning_rate,
-                    beta_1: beta_1,
-                    beta_2: beta_2,
-                    epsilon: epsilon,
-                    amsgrad: amsgrad,
-                    name: name);
-        }
-    }
-}

src/TensorFlowNET.Core/APIs/tf.train.cs

@@ -15,7 +15,6 @@
 ******************************************************************************/
 
 using System.Collections.Generic;
-using Tensorflow.Keras.Optimizers;
 using Tensorflow.Train;
 
 namespace Tensorflow
@@ -87,7 +86,7 @@ namespace Tensorflow
             public CheckpointState get_checkpoint_state(string checkpoint_dir, string latest_filename = null)
                 => checkpoint_management.get_checkpoint_state(checkpoint_dir, latest_filename: latest_filename);
 
-            public Tensor polynomial_decay(float learning_rate,
+            /*public Tensor polynomial_decay(float learning_rate,
                 RefVariable global_step,
                 float decay_steps,
                 float end_learning_rate = 0.0001f,
@@ -105,7 +104,7 @@ namespace Tensorflow
                 var decayed_lr = decayed.__call__(global_step);
 
                 return decayed_lr;
-            }
+            }*/
         }
     }
 }

src/TensorFlowNET.Core/Keras/ArgsDefinition/DataHandlerArgs.cs

@@ -14,7 +14,7 @@ namespace Tensorflow.Keras.ArgsDefinition
         public int MaxQueueSize { get; set; } = 10;
         public int Workers { get; set; } = 1;
         public bool UseMultiprocessing { get; set; } = false;
-        public Model Model { get; set; }
+        public IModel Model { get; set; }
         public IVariableV1 StepsPerExecution { get; set; }
     }
 }

src/TensorFlowNET.Core/Keras/ArgsDefinition/NodeArgs.cs

@@ -4,7 +4,7 @@ namespace Tensorflow.Keras.ArgsDefinition
 {
     public class NodeArgs
     {
-        public Layer[] InboundLayers { get; set; }
+        public ILayer[] InboundLayers { get; set; }
         public int[] NodeIndices { get; set; }
         public int[] TensorIndices { get; set; }
         public Tensors InputTensors { get; set; }

src/TensorFlowNET.Core/Keras/ArgsDefinition/SequentialArgs.cs

@@ -5,6 +5,6 @@ namespace Tensorflow.Keras.ArgsDefinition
 {
     public class SequentialArgs : ModelArgs
     {
-        public List<Layer> Layers { get; set; }
+        public List<ILayer> Layers { get; set; }
     }
 }

src/TensorFlowNET.Core/Keras/ArgsDefinition/TensorLikeDataAdapterArgs.cs

@@ -13,6 +13,6 @@ namespace Tensorflow.Keras.ArgsDefinition
         public int MaxQueueSize { get; set; }
         public int Worker { get; set; }
         public bool UseMultiprocessing { get; set; }
-        public Model Model { get; set; }
+        public IModel Model { get; set; }
     }
 }

src/TensorFlowNET.Core/Keras/Datasets/Mnist.cs

@@ -1,74 +0,0 @@
-/*****************************************************************************
-   Copyright 2020 Haiping Chen. All Rights Reserved.
-
-   Licensed under the Apache License, Version 2.0 (the "License");
-   you may not use this file except in compliance with the License.
-   You may obtain a copy of the License at
-
-       http://www.apache.org/licenses/LICENSE-2.0
-
-   Unless required by applicable law or agreed to in writing, software
-   distributed under the License is distributed on an "AS IS" BASIS,
-   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-   See the License for the specific language governing permissions and
-   limitations under the License.
-******************************************************************************/
-
-using NumSharp;
-using System;
-using System.IO;
-using System.Net;
-
-namespace Tensorflow.Keras.Datasets
-{
-    public class Mnist
-    {
-        string origin_folder = "https://storage.googleapis.com/tensorflow/tf-keras-datasets/";
-        string file_name = "mnist.npz";
-
-        /// <summary>
-        /// Loads the [MNIST dataset](http://yann.lecun.com/exdb/mnist/).
-        /// </summary>
-        /// <returns></returns>
-        public DatasetPass load_data()
-        {
-            var file = Download();
-            var bytes = File.ReadAllBytes(file);
-            var datax = LoadX(bytes);
-            var datay = LoadY(bytes);
-            return new DatasetPass
-            {
-                Train = (datax.Item1, datay.Item1),
-                Test = (datax.Item2, datay.Item2)
-            };
-        }
-
-        (NDArray, NDArray) LoadX(byte[] bytes)
-        {
-            var y = np.Load_Npz<byte[,,]>(bytes);
-            return (y["x_train.npy"], y["x_test.npy"]);
-        }
-
-        (NDArray, NDArray) LoadY(byte[] bytes)
-        {
-            var y = np.Load_Npz<byte[]>(bytes);
-            return (y["y_train.npy"], y["y_test.npy"]);
-        }
-
-        string Download()
-        {
-            var fileSaveTo = Path.Combine(Path.GetTempPath(), file_name);
-
-            if (File.Exists(fileSaveTo))
-            {
-                Console.WriteLine($"The file {fileSaveTo} already exists");
-                return fileSaveTo;
-            }
-
-            using var wc = new WebClient();
-            wc.DownloadFileTaskAsync(origin_folder + file_name, fileSaveTo).Wait();
-
-            return fileSaveTo;
-        }
-    }
-}

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

@@ -1,10 +0,0 @@
-namespace Tensorflow.Keras.Engine
-{
-    public class CallContext
-    {
-        public CallContextManager enter()
-        {
-            return new CallContextManager();
-        }
-    }
-}

src/TensorFlowNET.Keras/Engine/InputSpec.cs → src/TensorFlowNET.Core/Keras/Engine/IModel.cs

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

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

@@ -0,0 +1,16 @@
+using System;
+using System.Collections.Generic;
+using System.Text;
+
+namespace Tensorflow.Keras.Engine
+{
+    public interface INode
+    {
+        Tensors input_tensors { get; }
+        Tensors Outputs { get; }
+        ILayer Layer { get; set; }
+        List<Tensor> KerasInputs { get; set; }
+        INode[] ParentNodes { get; }
+        IEnumerable<(ILayer, int, int, Tensor)> iterate_inbound();
+    }
+}

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

@@ -5,12 +5,13 @@
     /// </summary>
     public class KerasHistory
     {
-        Layer layer;
+        ILayer layer;
+        public ILayer Layer => layer;
         int node_index;
         int tensor_index;
         Tensor tensor;
 
-        public KerasHistory(Layer layer, int node_index, int tensor_index, Tensor tensor)
+        public KerasHistory(ILayer layer, int node_index, int tensor_index, Tensor tensor)
         {
             this.layer = layer;
             this.node_index = node_index;
@@ -18,7 +19,7 @@
             this.tensor = tensor;
         }
 
-        public void Deconstruct(out Layer layer, out int node_index, out int tensor_index)
+        public void Deconstruct(out ILayer layer, out int node_index, out int tensor_index)
         {
             layer = this.layer;
             node_index = this.node_index;
@@ -27,8 +28,5 @@
 
         public override string ToString()
             => $"{layer.GetType().Name} {layer.Name} {tensor.name}";
-
-        public static implicit operator Layer(KerasHistory history)
-            => history.layer;
     }
 }

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

@@ -1,121 +0,0 @@
-/*****************************************************************************
-   Copyright 2018 The TensorFlow.NET Authors. All Rights Reserved.
-
-   Licensed under the Apache License, Version 2.0 (the "License");
-   you may not use this file except in compliance with the License.
-   You may obtain a copy of the License at
-
-       http://www.apache.org/licenses/LICENSE-2.0
-
-   Unless required by applicable law or agreed to in writing, software
-   distributed under the License is distributed on an "AS IS" BASIS,
-   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-   See the License for the specific language governing permissions and
-   limitations under the License.
-******************************************************************************/
-
-using System.Collections.Generic;
-using System.Linq;
-using Tensorflow.Keras.ArgsDefinition;
-using static Tensorflow.Binding;
-
-namespace Tensorflow.Keras.Engine
-{
-    /// <summary>
-    /// A `Node` describes the connectivity between two layers.
-    /// 
-    /// Each time a layer is connected to some new input,
-    /// a node is added to `layer._inbound_nodes`.
-    /// Each time the output of a layer is used by another layer,
-    /// a node is added to `layer._outbound_nodes`.
-    /// </summary>
-    public partial class Node
-    {
-        NodeArgs args;
-
-        public int[] node_indices;
-        public int[] tensor_indices;
-        public Tensors input_tensors => args.InputTensors;
-        public Tensors Outputs => args.Outputs;
-        public TensorShape[] input_shapes;
-        public TensorShape[] output_shapes;
-        public List<Tensor> KerasInputs = new List<Tensor>();
-        public Layer Layer { get; set; }
-        public bool IsInput => args.InputTensors == null;
-        public int[] FlatInputIds { get; set; }
-        public int[] FlatOutputIds { get; set; }
-        bool _single_positional_tensor_passed => KerasInputs.Count() == 1;
-        Dictionary<int, int> _keras_inputs_ids_and_indices = new Dictionary<int, int>();
-        public Node[] ParentNodes
-        {
-            get
-            {
-                var node_deps = new List<Node>();
-                foreach (var kt in KerasInputs)
-                {
-                    var (layer, node_index, _) = kt.KerasHistory;
-                    if (layer != null)
-                        node_deps.append(layer.InboundNodes[node_index]);
-                }
-                return node_deps.ToArray();
-            }
-        }
-
-        public Node(Layer layer, NodeArgs args)
-        {
-            this.args = args;
-            this.Layer = layer;
-
-            if (args.InputTensors != null)
-                KerasInputs.AddRange(args.InputTensors);
-
-            foreach (var (i, ele) in enumerate(KerasInputs))
-                _keras_inputs_ids_and_indices[i] = ele.GetHashCode();
-
-            // Wire up Node to Layers.
-            layer.InboundNodes.Add(this);
-            foreach (var kt in KerasInputs)
-            {
-                if (kt.KerasHistory == null)
-                    continue;
-                var (inbound_layer, _, _) = kt.KerasHistory;
-                if (inbound_layer != null)
-                    inbound_layer.OutboundNodes.Add(this);
-            }
-
-            // Set metadata on outputs.
-            var node_index = layer.InboundNodes.Count - 1;
-            foreach (var (i, tensor) in enumerate(Outputs))
-                tensor.KerasHistory = new KerasHistory(layer, node_index, i, tensor);
-
-            // Cached for performance.
-            FlatInputIds = KerasInputs.Select(x => x.GetHashCode()).ToArray();
-            FlatOutputIds = Outputs.Select(x => x.GetHashCode()).ToArray();
-        }
-
-        /// <summary>
-        /// Maps Keras Tensors to computed Tensors using `tensor_dict`.
-        /// </summary>
-        /// <param name="tensor_dict"></param>
-        /// <returns></returns>
-        public Tensors MapArguments(Dictionary<int, Queue<Tensor>> tensor_dict)
-        {
-            if (_single_positional_tensor_passed)
-            {
-                var kt_id = _keras_inputs_ids_and_indices[0];
-                return tensor_dict[kt_id].Dequeue();
-            }
-            else
-            {
-                var flat_arguments = KerasInputs.Select(x => x).ToArray();
-                foreach (var (kt_index, kt_id) in enumerate(_keras_inputs_ids_and_indices))
-                    flat_arguments[kt_index] = tensor_dict[kt_id].Dequeue();
-
-                return flat_arguments;
-            }
-        }
-
-        public override string ToString()
-            => $"{Layer.Name}, {KerasInputs.Count} inputs: {string.Join(",", KerasInputs.Select(x => x.name))}";
-    }
-}

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

@@ -1,134 +0,0 @@
-/*****************************************************************************
-   Copyright 2018 The TensorFlow.NET Authors. All Rights Reserved.
-
-   Licensed under the Apache License, Version 2.0 (the "License");
-   you may not use this file except in compliance with the License.
-   You may obtain a copy of the License at
-
-       http://www.apache.org/licenses/LICENSE-2.0
-
-   Unless required by applicable law or agreed to in writing, software
-   distributed under the License is distributed on an "AS IS" BASIS,
-   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-   See the License for the specific language governing permissions and
-   limitations under the License.
-******************************************************************************/
-
-using System.Collections.Generic;
-using Tensorflow.Keras.ArgsDefinition;
-using Tensorflow.Keras.Layers;
-using static Tensorflow.Binding;
-
-namespace Tensorflow.Keras.Engine
-{
-    /// <summary>
-    /// `Sequential` groups a linear stack of layers into a `tf.keras.Model`.
-    /// `Sequential` provides training and inference features on this model.
-    /// </summary>
-    public class Sequential : Model
-    {
-        SequentialArgs args;
-        bool _is_graph_network;
-        Tensor inputs;
-        Tensor outputs;
-
-        bool computeOutputAndMaskJointly;
-        bool autoTrackSubLayers;
-        TensorShape inferredInputShape;
-        bool hasExplicitInputShape;
-        TF_DataType inputDType;
-        List<Layer> layers => args.Layers;
-        public TensorShape output_shape => outputs.TensorShape;
-        bool built = false;
-
-        public Sequential(SequentialArgs args)
-            : base(new ModelArgs
-            {
-                Name = args.Name
-            })
-        {
-            this.args = args;
-            if (args.Layers == null)
-                args.Layers = new List<Layer>();
-            // SupportsMasking = true;
-            computeOutputAndMaskJointly = true;
-            autoTrackSubLayers = false;
-            hasExplicitInputShape = false;
-            _is_graph_network = false;
-        }
-
-        public void add(Tensor tensor)
-        {
-            Layer layer = tensor.KerasHistory;
-            add(layer);
-        }
-
-        /// <summary>
-        /// Adds a layer instance on top of the layer stack.
-        /// </summary>
-        /// <param name="layer"></param>
-        public void add(Layer layer)
-        {
-            built = false;
-            var set_inputs = false;
-            if (layers.Count == 0)
-            {
-                if (layer is InputLayer)
-                {
-                    set_inputs = true;
-                }
-                else
-                {
-                    if (layer.BatchInputShape != null)
-                    {
-                        // Instantiate an input layer.
-                        var x = tf.keras.Input(
-                              shape: layer.BatchInputShape,
-                              dtype: layer.DType,
-                              name: layer.Name + "_input");
-
-                        // This will build the current layer
-                        // and create the node connecting the current layer
-                        // to the input layer we just created.
-                        layer.Apply(x);
-                        set_inputs = true;
-                    }
-                }
-
-                if (set_inputs)
-                {
-                    // If an input layer (placeholder) is available.
-                    outputs = layer.InboundNodes[^1].Outputs;
-                }
-
-            }
-            else if (outputs != null)
-            {
-                outputs = layer.Apply(outputs);
-            }
-
-            if (set_inputs || _is_graph_network)
-            {
-                _init_graph_network(inputs, outputs);
-            }
-            else
-            {
-
-            }
-        }
-
-        void _init_graph_network(Tensor inputs, Tensor outputs)
-        {
-            _is_graph_network = true;
-            this.inputs = inputs;
-            this.outputs = outputs;
-            built = true;
-            _map_graph_network(inputs, outputs);
-        }
-
-        void _map_graph_network(Tensor inputs, Tensor outputs)
-        {
-            layers.add(outputs.KerasHistory);
-        }
-    }
-}

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

@@ -0,0 +1,20 @@
+using System;
+using System.Collections.Generic;
+using System.Text;
+using Tensorflow.Keras.Engine;
+
+namespace Tensorflow.Keras
+{
+    public interface ILayer
+    {
+        string Name { get; }
+        bool Trainable { get; }
+        List<ILayer> Layers { get; }
+        List<INode> InboundNodes { get; }
+        List<INode> OutboundNodes { get; }
+        Tensors Apply(Tensors inputs, Tensor state = null, bool is_training = false);
+        List<IVariableV1> trainable_variables { get; }
+        TensorShape output_shape { get; }
+        int count_params();
+    }
+}

src/TensorFlowNET.Core/Keras/Losses/Loss.cs

@@ -1,41 +0,0 @@
-using System;
-using Tensorflow.Keras.Utils;
-
-namespace Tensorflow.Keras.Losses
-{
-    /// <summary>
-    /// Loss base class.
-    /// </summary>
-    public abstract class Loss
-    {
-        protected string reduction;
-        protected string name;
-        bool _allow_sum_over_batch_size;
-        string _name_scope;
-
-        public string Reduction => reduction;
-
-        public Loss(string reduction = ReductionV2.AUTO, string name = null)
-        {
-            this.reduction = reduction;
-            this.name = name;
-            _allow_sum_over_batch_size = false;
-        }
-
-        public virtual Tensor Apply(Tensor y_true, Tensor y_pred, bool from_logits = false, int axis = -1)
-        {
-            throw new NotImplementedException("");
-        }
-
-        public Tensor Call(Tensor y_true, Tensor y_pred)
-        {
-            var losses = Apply(y_true, y_pred);
-            return losses_utils.compute_weighted_loss(losses, reduction: ReductionV2.SUM_OVER_BATCH_SIZE);
-        }
-
-        void _set_name_scope()
-        {
-            _name_scope = name;
-        }
-    }
-}

src/TensorFlowNET.Core/Keras/Losses/LossFunctionWrapper.cs

@@ -1,12 +0,0 @@
-namespace Tensorflow.Keras.Losses
-{
-    public class LossFunctionWrapper : Loss
-    {
-        public LossFunctionWrapper(string reduction = ReductionV2.AUTO,
-            string name = null)
-            : base(reduction: reduction,
-                  name: name)
-        {
-        }
-    }
-}

src/TensorFlowNET.Core/Keras/Losses/SparseCategoricalCrossentropy.cs

@@ -1,33 +0,0 @@
-using static Tensorflow.Binding;
-
-namespace Tensorflow.Keras.Losses
-{
-    public class SparseCategoricalCrossentropy : LossFunctionWrapper, ILossFunc
-    {
-        public SparseCategoricalCrossentropy(bool from_logits = false,
-            string reduction = ReductionV2.AUTO,
-            string name = "sparse_categorical_crossentropy") :
-            base(reduction: reduction,
-                name: name)
-        {
-
-        }
-
-        public override Tensor Apply(Tensor target, Tensor output, bool from_logits = false, int axis = -1)
-        {
-            target = tf.cast(target, dtype: TF_DataType.TF_INT64);
-
-            // Try to adjust the shape so that rank of labels = rank of logits - 1.
-            var output_shape = array_ops.shape_v2(output);
-            var output_rank = output.TensorShape.ndim;
-            var target_rank = target.TensorShape.ndim;
-            var update_shape = target_rank != output_rank - 1;
-            if (update_shape)
-            {
-                target = array_ops.reshape(target, new int[] { -1 });
-                output = array_ops.reshape(output, new int[] { -1, output_shape[-1].numpy() });
-            }
-            return tf.nn.sparse_softmax_cross_entropy_with_logits(target, output);
-        }
-    }
-}

src/TensorFlowNET.Core/Keras/Metrics/Mean.cs

@@ -1,14 +0,0 @@
-namespace Tensorflow.Keras.Metrics
-{
-    /// <summary>
-    /// Computes the (weighted) mean of the given values.
-    /// </summary>
-    public class Mean : Reduce
-    {
-        public Mean(string name = "mean", TF_DataType dtype = TF_DataType.TF_FLOAT)
-            : base(Reduction.WEIGHTED_MEAN, name, dtype: dtype)
-        {
-
-        }
-    }
-}

src/TensorFlowNET.Core/Keras/Metrics/MeanMetricWrapper.cs

@@ -1,27 +0,0 @@
-using System;
-using System.Collections.Generic;
-using System.Text;
-
-namespace Tensorflow.Keras.Metrics
-{
-    public class MeanMetricWrapper : Mean
-    {
-        string name;
-        Func<Tensor, Tensor, Tensor> _fn = null;
-        
-        public MeanMetricWrapper(Func<Tensor, Tensor, Tensor> fn, string name, TF_DataType dtype = TF_DataType.TF_FLOAT)
-            : base(name: name, dtype: dtype)
-        {
-            _fn = fn;
-        }
-
-        public override Tensor update_state(Tensor y_true, Tensor y_pred, Tensor sample_weight = null)
-        {
-            y_true = math_ops.cast(y_true, _dtype);
-            y_pred = math_ops.cast(y_pred, _dtype);
-
-            var matches = _fn(y_true, y_pred);
-            return update_state(matches, sample_weight: sample_weight);
-        }
-    }
-}

src/TensorFlowNET.Core/Keras/Metrics/Metric.cs

@@ -1,62 +0,0 @@
-using System;
-using Tensorflow.Keras.ArgsDefinition;
-using Tensorflow.Keras.Engine;
-using static Tensorflow.Binding;
-
-namespace Tensorflow.Keras.Metrics
-{
-    /// <summary>
-    /// Encapsulates metric logic and state.
-    /// </summary>
-    public class Metric : Layer
-    {
-        protected IVariableV1 total;
-        protected IVariableV1 count;
-        protected string _reduction;
-        protected TF_DataType _dtype;
-
-        public Metric(string name = null, TF_DataType dtype = TF_DataType.DtInvalid)
-            : base(new LayerArgs
-            {
-                Name = name,
-                DType = dtype
-            })
-        {
-            stateful = true;
-            built = true;
-        }
-
-        protected override IVariableV1 add_weight(string name,
-            TensorShape shape = null,
-            TF_DataType dtype = TF_DataType.TF_FLOAT,
-            IInitializer initializer = null,
-            IRegularizer regularizer = null,
-            VariableSynchronization synchronization = VariableSynchronization.OnRead,
-            VariableAggregation aggregation = VariableAggregation.Sum,
-            bool trainable = true,
-            Func<VariableArgs, IVariableV1> getter = null)
-        {
-            if (shape == null)
-                shape = new TensorShape(new int[0]);
-
-            return tf_with(ops.init_scope(), delegate
-            {
-                return base.add_weight(name, shape,
-                    dtype: dtype,
-                    trainable: false,
-                    initializer: initializer,
-                    synchronization: synchronization,
-                    aggregation: aggregation);
-            });
-        }
-
-        public virtual Tensor update_state(Tensor y_true, Tensor y_pred, Tensor sample_weight = null)
-            => throw new NotImplementedException("");
-
-        public virtual Tensor result()
-            => throw new NotImplementedException("");
-
-        public override string ToString()
-            => $"{name} {(float)total.numpy()}/{(float)count.numpy()}";
-    }
-}

src/TensorFlowNET.Core/Keras/Metrics/Reduce.cs

@@ -1,74 +0,0 @@
-using Tensorflow.Keras.Losses;
-using Tensorflow.Keras.Utils;
-using static Tensorflow.Binding;
-
-namespace Tensorflow.Keras.Metrics
-{
-    /// <summary>
-    /// Encapsulates metrics that perform a reduce operation on the values.
-    /// </summary>
-    public class Reduce : Metric
-    {
-        public Reduce(string reduction, string name, TF_DataType dtype = TF_DataType.DtInvalid)
-            : base(name: name, dtype: dtype)
-        {
-            _reduction = reduction;
-            _dtype = dtype;
-            total = add_weight("total", initializer: tf.zeros_initializer);
-
-            if (reduction == Reduction.WEIGHTED_MEAN ||
-                reduction == Reduction.SUM_OVER_BATCH_SIZE)
-            {
-                count = add_weight("count", initializer: tf.zeros_initializer);
-            }
-        }
-
-        public Tensor update_state(Tensor values, Tensor sample_weight = null)
-        {
-            if (sample_weight != null)
-            {
-                (values, sample_weight) = losses_utils.squeeze_or_expand_dimensions(
-                    values, sample_weight: sample_weight);
-
-                sample_weight = math_ops.cast(sample_weight, dtype: values.dtype);
-                values = math_ops.multiply(values, sample_weight);
-            }
-
-            Tensor update_total_op = null;
-            var value_sum = math_ops.reduce_sum(values);
-            tf_with(ops.control_dependencies(new[] { value_sum }), ctl =>
-            {
-                update_total_op = total.assign_add(value_sum);
-            });
-
-            // Exit early if the reduction doesn't have a denominator.
-            if (_reduction == Reduction.SUM)
-                return update_total_op;
-
-            // Update `count` for reductions that require a denominator.
-            Tensor num_values = null;
-            if (_reduction == Reduction.SUM_OVER_BATCH_SIZE)
-                num_values = math_ops.cast(array_ops.size(values), _dtype);
-            else if (_reduction == ReductionV2.WEIGHTED_MEAN)
-            {
-                if (sample_weight == null)
-                    num_values = math_ops.cast(array_ops.size(values), _dtype);
-                else
-                    num_values = math_ops.reduce_sum(sample_weight);
-            }
-
-            return tf_with(ops.control_dependencies(new[] { update_total_op }), ctl
-                => count.assign_add(num_values));
-        }
-
-        public override Tensor result()
-        {
-            if (_reduction == Reduction.SUM)
-                return array_ops.identity(total.AsTensor());
-            else if (_reduction == Reduction.WEIGHTED_MEAN || _reduction == Reduction.SUM_OVER_BATCH_SIZE)
-                return math_ops.div_no_nan(total.AsTensor(), count.AsTensor());
-
-            return base.result();
-        }
-    }
-}

src/TensorFlowNET.Core/Keras/Metrics/Sum.cs

@@ -1,6 +0,0 @@
-namespace Tensorflow.Keras.Metrics
-{
-    class Sum
-    {
-    }
-}

src/TensorFlowNET.Core/Layers/Layer.cs

@@ -1,210 +0,0 @@
-/*****************************************************************************
-   Copyright 2018 The TensorFlow.NET Authors. All Rights Reserved.
-
-   Licensed under the Apache License, Version 2.0 (the "License");
-   you may not use this file except in compliance with the License.
-   You may obtain a copy of the License at
-
-       http://www.apache.org/licenses/LICENSE-2.0
-
-   Unless required by applicable law or agreed to in writing, software
-   distributed under the License is distributed on an "AS IS" BASIS,
-   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
-   See the License for the specific language governing permissions and
-   limitations under the License.
-******************************************************************************/
-
-using System;
-using System.Collections.Generic;
-using Tensorflow.Keras.ArgsDefinition;
-using static Tensorflow.Binding;
-
-namespace Tensorflow.Layers
-{
-    public class Layer : Keras.Engine.Layer
-    {
-        protected Graph _graph;
-
-        protected VariableScope _scope;
-        protected VariableScope _current_scope;
-
-        protected bool? _reuse;
-        protected bool _use_resource_variables;
-        protected bool _keras_style;
-
-        public Layer(bool trainable = true,
-            string name = null,
-            TF_DataType dtype = TF_DataType.DtInvalid,
-            bool? _reuse = null) :
-            base(new LayerArgs
-            {
-                Trainable = trainable,
-                Name = name,
-                DType = dtype
-            })
-        {
-            // For backwards compatibility, legacy layers do not use `ResourceVariable`
-            // by default.
-            this._use_resource_variables = false;
-            this._reuse = _reuse;
-
-            // Avoid an incorrect lint error
-            trainable_weights = new List<IVariableV1>();
-            non_trainable_weights = new List<IVariableV1>();
-            this.built = false;
-            _keras_style = false;
-        }
-
-        public virtual (Tensor, Tensor) apply(Tensor inputs, Tensor training = null)
-        {
-            var results = __call__(inputs, training: training);
-            return (results[0], results[1]);
-        }
-
-        public Tensors __call__(Tensors inputs,
-            Tensor state = null,
-            Tensor training = null,
-            VariableScope scope = null)
-        {
-            _set_scope(scope);
-            _graph = ops._get_graph_from_inputs(inputs, graph: _graph);
-
-            variable_scope scope_context_manager = null;
-            if (built)
-            {
-                scope_context_manager = tf.variable_scope(_scope,
-                    reuse: true,
-                    auxiliary_name_scope: false);
-            }
-            else
-            {
-                scope_context_manager = tf.variable_scope(_scope,
-                    reuse: _reuse,
-                    auxiliary_name_scope: false);
-            }
-
-            Tensors outputs = null;
-            tf_with(scope_context_manager, scope2 =>
-            {
-                _current_scope = scope2;
-                // Actually call layer
-                outputs = base.Apply(inputs,
-                    state: state,
-                    is_training: training == null ? false : false);
-            });
-
-
-            // Update global default collections.
-            _add_elements_to_collection(updates.ToArray(), new string[] { tf.GraphKeys.UPDATE_OPS });
-
-            return outputs;
-        }
-
-        protected virtual void _add_elements_to_collection(Operation[] elements, string[] collection_list)
-        {
-            foreach (var name in collection_list)
-            {
-                var collection = ops.get_collection_ref<Operation>(name);
-
-                foreach (var element in elements)
-                    if (!collection.Contains(element))
-                        collection.Add(element);
-            }
-        }
-
-        /// <summary>
-        /// Adds a new variable to the layer, or gets an existing one; returns it.
-        /// </summary>
-        /// <param name="name"></param>
-        /// <param name="shape"></param>
-        /// <param name="dtype"></param>
-        /// <param name="initializer"></param>
-        /// <param name="trainable"></param>
-        /// <param name="synchronization"></param>
-        /// <param name="aggregation"></param>
-        /// <returns></returns>
-        protected virtual IVariableV1 add_weight(string name,
-            int[] shape,
-            TF_DataType dtype = TF_DataType.DtInvalid,
-            IInitializer initializer = null,
-            bool trainable = true,
-            VariableSynchronization synchronization = VariableSynchronization.Auto,
-            VariableAggregation aggregation = VariableAggregation.None)
-        {
-            var default_graph = ops.get_default_graph();
-            Graph init_graph = null;
-            IVariableV1[] existing_variables = null;
-
-            if (synchronization == VariableSynchronization.OnRead)
-                trainable = false;
-
-            if (default_graph.building_function)
-            {
-                throw new NotImplementedException("add_weight");
-            }
-            else
-            {
-                init_graph = default_graph;
-                existing_variables = variables.global_variables().ToArray();
-            }
-
-            if (dtype == TF_DataType.DtInvalid)
-                dtype = TF_DataType.TF_FLOAT;
-
-            _set_scope();
-            var reuse = built || (_reuse != null && _reuse.Value);
-            return tf_with(tf.variable_scope(_scope,
-                reuse: reuse,
-                auxiliary_name_scope: false), scope =>
-                {
-                    _current_scope = scope;
-                    return tf_with(ops.name_scope(_name_scope()), delegate
-                    {
-                        var variable = base.add_weight(name,
-                            shape,
-                            dtype: dtype,
-                            initializer: initializer,
-                            trainable: trainable,
-                            getter: (args) =>
-                                tf.compat.v1.get_variable(args.Name,
-                                    shape: args.Shape,
-                                    dtype: args.DType,
-                                    initializer: args.Initializer,
-                                    trainable: args.Trainable)
-                            );
-
-                        //if (init_graph != null)
-                        //var trainable_variables = variables.trainable_variables();
-
-                        return variable;
-                    });
-                });
-        }
-
-        protected override string _name_scope()
-        {
-            return _current_scope.original_name_scope;
-        }
-
-        protected void _set_scope(VariableScope scope = null)
-        {
-            if (_scope == null)
-            {
-                if (_reuse.HasValue && _reuse.Value)
-                {
-                    throw new NotImplementedException("_set_scope _reuse.HasValue");
-                    /*with(tf.variable_scope(scope == null ? _base_name : scope),
-                        captured_scope => _scope = captured_scope);*/
-                }
-                else
-                {
-                    tf_with(tf.variable_scope(scope, default_name: base_name), captured_scope =>
-                    {
-                        // convert variable_scope to VariableScope
-                        _scope = captured_scope;
-                    });
-                }
-            }
-        }
-    }
-}

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

@@ -71,7 +71,7 @@ namespace Tensorflow
         /// <param name="training"></param>
         /// <param name="state"></param>
         /// <returns></returns>
-        protected override Tensors Call(Tensors inputs, Tensor state = null, bool is_training = false)
+        protected Tensors Call(Tensors inputs, Tensor state = null, bool is_training = false)
         {
             var one = constant_op.constant(1, dtype: dtypes.int32);
             // Parameters of gates are concatenated into one multiply for efficiency.

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

@@ -66,7 +66,7 @@ namespace Tensorflow
             built = true;
         }
 
-        protected override Tensors Call(Tensors inputs, Tensor state = null, bool is_training = false)
+        protected Tensors Call(Tensors inputs, Tensor state = null, bool is_training = false)
         {
             // Most basic RNN: output = new_state = act(W * input + U * state + B).
             var concat = array_ops.concat(new Tensor[] { inputs, state }, 1);

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

@@ -22,7 +22,7 @@ using static Tensorflow.Binding;
 
 namespace Tensorflow.Operations
 {
-    internal class ConvolutionInternal
+    public class ConvolutionInternal
     {
         ConvolutionalArgs args;
 

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

@@ -13,17 +13,165 @@
    See the License for the specific language governing permissions and
    limitations under the License.
 ******************************************************************************/
+using static Tensorflow.Binding;
+using Tensorflow.Keras.Engine;
+using System;
 
 namespace Tensorflow
 {
     public class LayerRnnCell : RnnCell
     {
-        public LayerRnnCell(bool? _reuse = null,
-            string name = null,
-            TF_DataType dtype = TF_DataType.DtInvalid) : base(_reuse: _reuse,
+        protected InputSpec inputSpec;
+        protected bool built;
+        protected Graph _graph;
+
+        protected VariableScope _scope;
+        protected VariableScope _current_scope;
+
+        protected bool? _reuse;
+        protected bool _use_resource_variables;
+        protected bool _keras_style;
+
+        public LayerRnnCell(bool trainable = true,
+                    string name = null,
+                    TF_DataType dtype = TF_DataType.DtInvalid,
+                    bool? _reuse = null) : base(_reuse: _reuse,
                     name: name,
                     dtype: dtype)
         {
+            // For backwards compatibility, legacy layers do not use `ResourceVariable`
+            // by default.
+            this._use_resource_variables = false;
+            this._reuse = _reuse;
+
+            // Avoid an incorrect lint error
+            this.built = false;
+            _keras_style = false;
+        }
+
+        protected virtual void build(TensorShape inputs_shape)
+        {
+
+        }
+
+        public virtual (Tensor, Tensor) apply(Tensor inputs, Tensor training = null)
+        {
+            var results = __call__(inputs, training: training);
+            return (results[0], results[1]);
+        }
+
+        public Tensors __call__(Tensors inputs,
+            Tensor state = null,
+            Tensor training = null,
+            VariableScope scope = null)
+        {
+            _set_scope(scope);
+            _graph = ops._get_graph_from_inputs(inputs, graph: _graph);
+
+            variable_scope scope_context_manager = null;
+            if (built)
+            {
+                scope_context_manager = tf.variable_scope(_scope,
+                    reuse: true,
+                    auxiliary_name_scope: false);
+            }
+            else
+            {
+                scope_context_manager = tf.variable_scope(_scope,
+                    reuse: _reuse,
+                    auxiliary_name_scope: false);
+            }
+
+            Tensors outputs = null;
+            tf_with(scope_context_manager, scope2 =>
+            {
+                _current_scope = scope2;
+                // Actually call layer
+
+            });
+
+
+            // Update global default collections.
+
+            return outputs;
+        }
+
+        protected virtual void _add_elements_to_collection(Operation[] elements, string[] collection_list)
+        {
+            foreach (var name in collection_list)
+            {
+                var collection = ops.get_collection_ref<Operation>(name);
+
+                foreach (var element in elements)
+                    if (!collection.Contains(element))
+                        collection.Add(element);
+            }
+        }
+
+        /// <summary>
+        /// Adds a new variable to the layer, or gets an existing one; returns it.
+        /// </summary>
+        /// <param name="name"></param>
+        /// <param name="shape"></param>
+        /// <param name="dtype"></param>
+        /// <param name="initializer"></param>
+        /// <param name="trainable"></param>
+        /// <param name="synchronization"></param>
+        /// <param name="aggregation"></param>
+        /// <returns></returns>
+        protected virtual IVariableV1 add_weight(string name,
+            int[] shape,
+            TF_DataType dtype = TF_DataType.DtInvalid,
+            IInitializer initializer = null,
+            bool trainable = true,
+            VariableSynchronization synchronization = VariableSynchronization.Auto,
+            VariableAggregation aggregation = VariableAggregation.None)
+        {
+            var default_graph = ops.get_default_graph();
+            Graph init_graph = null;
+            IVariableV1[] existing_variables = null;
+
+            if (synchronization == VariableSynchronization.OnRead)
+                trainable = false;
+
+            if (default_graph.building_function)
+            {
+                throw new NotImplementedException("add_weight");
+            }
+            else
+            {
+                init_graph = default_graph;
+                existing_variables = variables.global_variables().ToArray();
+            }
+
+            if (dtype == TF_DataType.DtInvalid)
+                dtype = TF_DataType.TF_FLOAT;
+
+            _set_scope();
+            var reuse = built || (_reuse != null && _reuse.Value);
+            return tf.Variable(0);
+        }
+
+        protected string _name_scope()
+        {
+            return _current_scope.original_name_scope;
+        }
+
+        protected void _set_scope(VariableScope scope = null)
+        {
+            if (_scope == null)
+            {
+                if (_reuse.HasValue && _reuse.Value)
+                {
+                    throw new NotImplementedException("_set_scope _reuse.HasValue");
+                    /*with(tf.variable_scope(scope == null ? _base_name : scope),
+                        captured_scope => _scope = captured_scope);*/
+                }
+                else
+                {
+
+                }
+            }
         }
     }
 }

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

@@ -15,6 +15,9 @@
 ******************************************************************************/
 
 using System;
+using System.Collections.Generic;
+using Tensorflow.Keras;
+using Tensorflow.Keras.Engine;
 using Tensorflow.Operations;
 using Tensorflow.Util;
 using static Tensorflow.Binding;
@@ -42,7 +45,7 @@ namespace Tensorflow
     /// matching structure of Tensors having shape `[batch_size].concatenate(s)`
     /// for each `s` in `self.batch_size`.
     /// </summary>
-    public abstract class RnnCell : Layers.Layer
+    public abstract class RnnCell : ILayer
     {
         /// <summary>
         /// Attribute that indicates whether the cell is a TF RNN cell, due the slight
@@ -52,14 +55,24 @@ namespace Tensorflow
         public virtual object state_size { get; }
 
         public virtual int output_size { get; }
+        public string Name { get => throw new NotImplementedException(); set => throw new NotImplementedException(); }
+
+        public List<INode> InboundNodes => throw new NotImplementedException();
+
+        public List<INode> OutboundNodes => throw new NotImplementedException();
+
+        public List<ILayer> Layers => throw new NotImplementedException();
+
+        public bool Trainable => throw new NotImplementedException();
+
+        public List<IVariableV1> trainable_variables => throw new NotImplementedException();
+
+        public TensorShape output_shape => throw new NotImplementedException();
 
         public RnnCell(bool trainable = true,
             string name = null,
             TF_DataType dtype = TF_DataType.DtInvalid,
-            bool? _reuse = null) : base(trainable: trainable,
-                    name: name,
-                    dtype: dtype,
-                    _reuse: _reuse)
+            bool? _reuse = null)
         {
             _is_tf_rnn_cell = true;
         }
@@ -109,5 +122,15 @@ namespace Tensorflow
 
             throw new NotImplementedException("_zero_state_tensors");
         }
+
+        public Tensors Apply(Tensors inputs, Tensor state = null, bool is_training = false)
+        {
+            throw new NotImplementedException();
+        }
+
+        public int count_params()
+        {
+            throw new NotImplementedException();
+        }
     }
 }

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

@@ -363,8 +363,8 @@ namespace Tensorflow.Operations
                 Tensor[] outputs = null;
                 if (sequence_length != null)
                     throw new NotImplementedException("sequence_length != null");
-                else
-                    outputs = cell.__call__(input_t_t, state: state1);
+                /*else
+                    outputs = cell.__call__(input_t_t, state: state1);*/
 
                 var (output, new_state) = (outputs[0], outputs[1]);
                 // Keras cells always wrap state as list, even if it's a single tensor.

src/TensorFlowNET.Core/Operations/nn_ops.cs

@@ -24,7 +24,7 @@ namespace Tensorflow
 {
     public class nn_ops
     {
-        internal static ConvolutionInternal convolution_internal(string padding,
+        public static ConvolutionInternal convolution_internal(string padding,
             int[] strides,
             int[] dilation_rate,
             string name = null,

src/TensorFlowNET.Core/Tensorflow.Binding.csproj

@@ -87,8 +87,4 @@ TensorFlow .NET v0.30 is focused on making more Keras API work including:
     <PackageReference Include="NumSharp.Lite" Version="0.1.9" />
     <PackageReference Include="Protobuf.Text" Version="0.4.0" />
   </ItemGroup>
-
-  <ItemGroup>
-    <Folder Include="Keras\Initializers\" />
-  </ItemGroup>
 </Project>

src/TensorFlowNET.Keras/Activations.cs

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

src/TensorFlowNET.Keras/Applications/Densenet.cs

@@ -1,35 +0,0 @@
-using System;
-using System.Collections.Generic;
-using System.Text;
-
-namespace Tensorflow.Keras.Applications
-{
-    public class Densenet
-    {
-        public static Tensor dense_block(Tensor x, int blocks, string name) => throw new NotImplementedException();
-
-        public static Tensor transition_block(Tensor x, float reduction, string name) => throw new NotImplementedException();
-
-        public static Tensor conv_block(Tensor x, float growth_rate, string name) => throw new NotImplementedException();
-
-        public static Model DenseNet(int blocks, bool include_top=true, string weights = "imagenet",
-                                    Tensor input_tensor = null, TensorShape input_shape = null, 
-                                    string pooling = null, int classes = 1000) => throw new NotImplementedException();
-
-        public static Model DenseNet121(int blocks, bool include_top = true, string weights = "imagenet",
-                                    Tensor input_tensor = null, TensorShape input_shape = null,
-                                    string pooling = null, int classes = 1000) => throw new NotImplementedException();
-
-        public static Model DenseNet169(int blocks, bool include_top = true, string weights = "imagenet",
-                                   Tensor input_tensor = null, TensorShape input_shape = null,
-                                   string pooling = null, int classes = 1000) => throw new NotImplementedException();
-
-        public static Model DenseNet201(int blocks, bool include_top = true, string weights = "imagenet",
-                                   Tensor input_tensor = null, TensorShape input_shape = null,
-                                   string pooling = null, int classes = 1000) => throw new NotImplementedException();
-
-        public static Tensor preprocess_input(Tensor x, string data_format = null) => throw new NotImplementedException();
-
-        public static Tensor decode_predictions(Tensor preds, int top = 5) => throw new NotImplementedException();
-    }
-}

src/TensorFlowNET.Keras/Applications/Efficientnet.cs

@@ -1,60 +0,0 @@
-using System;
-using System.Collections.Generic;
-using System.Text;
-
-namespace Tensorflow.Keras.Applications
-{
-    public class BlockArg
-    {
-
-    }
-
-    public class Efficientnet
-    {
-        public static Model EfficientNet(float width_coefficient, float depth_coefficient, int default_size, float dropout_rate = 0.2f,
-                                        float drop_connect_rate = 0.2f, int depth_divisor = 8, string activation = "swish",
-                                        BlockArg[] blocks_args = null, string model_name = "efficientnet", bool include_top = true,
-                                        string weights = "imagenet", Tensor input_tensor = null, TensorShape input_shape = null, 
-                                        string pooling = null, int classes = 1000) => throw new NotImplementedException();
-
-        public static Tensor block(Tensor inputs, string activation= "swish", float drop_rate= 0f,string name= "",
-                                    int filters_in= 32, int filters_out= 16, int kernel_size= 3, int strides= 1,
-                                    int expand_ratio= 1, float se_ratio= 0, bool  id_skip= true) => throw new NotImplementedException();
-
-        public static Model EfficientNetB0(bool include_top = true, string weights = "imagenet",
-                                    Tensor input_tensor = null, TensorShape input_shape = null,
-                                    string pooling = null, int classes = 1000) => throw new NotImplementedException();
-
-        public static Model EfficientNetB1(bool include_top = true, string weights = "imagenet",
-                                  Tensor input_tensor = null, TensorShape input_shape = null,
-                                  string pooling = null, int classes = 1000) => throw new NotImplementedException();
-
-        public static Model EfficientNetB2(bool include_top = true, string weights = "imagenet",
-                                  Tensor input_tensor = null, TensorShape input_shape = null,
-                                  string pooling = null, int classes = 1000) => throw new NotImplementedException();
-
-        public static Model EfficientNetB3(bool include_top = true, string weights = "imagenet",
-                                  Tensor input_tensor = null, TensorShape input_shape = null,
-                                  string pooling = null, int classes = 1000) => throw new NotImplementedException();
-
-        public static Model EfficientNetB4(bool include_top = true, string weights = "imagenet",
-                                 Tensor input_tensor = null, TensorShape input_shape = null,
-                                 string pooling = null, int classes = 1000) => throw new NotImplementedException();
-
-        public static Model EfficientNetB5(bool include_top = true, string weights = "imagenet",
-                                 Tensor input_tensor = null, TensorShape input_shape = null,
-                                 string pooling = null, int classes = 1000) => throw new NotImplementedException();
-
-        public static Model EfficientNetB6(bool include_top = true, string weights = "imagenet",
-                                 Tensor input_tensor = null, TensorShape input_shape = null,
-                                 string pooling = null, int classes = 1000) => throw new NotImplementedException();
-
-        public static Model EfficientNetB7(bool include_top = true, string weights = "imagenet",
-                                 Tensor input_tensor = null, TensorShape input_shape = null,
-                                 string pooling = null, int classes = 1000) => throw new NotImplementedException();
-
-        public static Tensor preprocess_input(Tensor x, string data_format = null) => throw new NotImplementedException();
-
-        public static Tensor decode_predictions(Tensor preds, int top = 5) => throw new NotImplementedException();
-    }
-}

src/TensorFlowNET.Keras/Applications/ImagenetUtils.cs

@@ -1,22 +0,0 @@
-using System;
-using System.Collections.Generic;
-using System.Text;
-
-namespace Tensorflow.Keras.Applications
-{
-    public class ImagenetUtils
-    {
-        public static Tensor preprocess_input(Tensor x, string data_format= null, string mode= "caffe") => throw new NotImplementedException();
-        
-        public static Tensor decode_predictions(Tensor preds, int top= 5) => throw new NotImplementedException();
-
-        public static Tensor _preprocess_numpy_input(Tensor x, string data_format, string mode) => throw new NotImplementedException();
-
-        public static Tensor _preprocess_symbolic_input(Tensor x, string data_format, string mode) => throw new NotImplementedException();
-
-        public static TensorShape obtain_input_shape(TensorShape input_shape, int default_size, int min_size,
-                       string data_format, bool require_flatten, string weights= null) => throw new NotImplementedException();
-
-        public static ((int, int), (int, int)) correct_pad(Tensor inputs, (int, int) kernel_size) => throw new NotImplementedException();
-    }
-}

src/TensorFlowNET.Keras/Applications/InceptionResnetV2.cs

@@ -1,22 +0,0 @@
-using System;
-using System.Collections.Generic;
-using System.Text;
-
-namespace Tensorflow.Keras.Applications
-{
-    public class InceptionResnetV2
-    {
-        public static Model InceptionResNetV2(bool include_top = true, string weights = "imagenet",
-                                   Tensor input_tensor = null, TensorShape input_shape = null,
-                                   string pooling = null, int classes = 1000) => throw new NotImplementedException();
-
-        public static Tensor conv2d_bn(Tensor x, int filters, (int, int) kernel_size, (int, int) strides, string padding= "same",
-                                        string activation= "relu", bool use_bias= false, string name= null) => throw new NotImplementedException();
-
-        public static Tensor inception_resnet_block(Tensor x, float scale, string block_type, int block_idx, string activation= "relu") => throw new NotImplementedException();
-
-        public static Tensor preprocess_input(Tensor x, string data_format = null) => throw new NotImplementedException();
-
-        public static Tensor decode_predictions(Tensor preds, int top = 5) => throw new NotImplementedException();
-    }
-}

src/TensorFlowNET.Keras/Applications/InceptionV3.cs

@@ -1,19 +0,0 @@
-using System;
-using System.Collections.Generic;
-using System.Text;
-
-namespace Tensorflow.Keras.Applications
-{
-    public class InceptionV3
-    {
-        public static Model Inceptionv3(bool include_top = true, string weights = "imagenet",
-                                  Tensor input_tensor = null, TensorShape input_shape = null,
-                                  string pooling = null, int classes = 1000) => throw new NotImplementedException();
-
-        public static Tensor conv2d_bn(Tensor x, int filters, int num_row, int num_col, string padding = "same", (int, int)? strides = null, string name = null) => throw new NotImplementedException();
-
-        public static Tensor preprocess_input(Tensor x, string data_format = null) => throw new NotImplementedException();
-
-        public static Tensor decode_predictions(Tensor preds, int top = 5) => throw new NotImplementedException();
-    }
-}

src/TensorFlowNET.Keras/Applications/Mobilenet.cs

@@ -1,18 +0,0 @@
-using System;
-using System.Collections.Generic;
-using System.Text;
-
-namespace Tensorflow.Keras.Applications
-{
-    public class Mobilenet
-    {
-        public static Model MobileNet(TensorShape input_shape= null, float alpha= 1.0f, int depth_multiplier= 1, float dropout= 1e-3f,
-                                    bool include_top= true, string weights= "imagenet", Tensor input_tensor= null, string pooling= null, int classes= 1000) => throw new NotImplementedException();
-
-        public static Tensor conv2d_bn(Tensor x, int filters, float alpha, (int, int)? kernel = null, (int, int)? strides = null) => throw new NotImplementedException();
-
-        public static Tensor preprocess_input(Tensor x, string data_format = null) => throw new NotImplementedException();
-
-        public static Tensor decode_predictions(Tensor preds, int top = 5) => throw new NotImplementedException();
-    }
-}

src/TensorFlowNET.Keras/Applications/MobilenetV2.cs

@@ -1,21 +0,0 @@
-using System;
-using System.Collections.Generic;
-using System.Text;
-
-namespace Tensorflow.Keras.Applications
-{
-    public class MobilenetV2
-    {
-        public static Model MobileNetV2(TensorShape input_shape = null, float alpha = 1.0f, bool include_top = true, 
-                                        string weights = "imagenet", Tensor input_tensor = null, string pooling = null,
-                                        int classes = 1000) => throw new NotImplementedException();
-
-        public static Tensor _inverted_res_block(Tensor inputs, int expansion, (int, int) stride, float alpha, int filters, string block_id) => throw new NotImplementedException();
-
-        public static Tensor _make_divisible(Tensor v, Tensor divisor, Tensor min_value= null) => throw new NotImplementedException();
-
-        public static Tensor preprocess_input(Tensor x, string data_format = null) => throw new NotImplementedException();
-
-        public static Tensor decode_predictions(Tensor preds, int top = 5) => throw new NotImplementedException();
-    }
-}

src/TensorFlowNET.Keras/Applications/Nasnet.cs

@@ -1,31 +0,0 @@
-using System;
-using System.Collections.Generic;
-using System.Text;
-
-namespace Tensorflow.Keras.Applications
-{
-    public class Nasnet
-    {
-        public static Model NASNet(TensorShape input_shape = null, int penultimate_filters = 4032, int num_blocks = 6, int stem_block_filters = 96,
-                                bool skip_reduction = true, int filter_multiplier = 2, bool include_top = true, string weights = null,
-                                Tensor input_tensor = null, string pooling = null, int classes = 1000, int? default_size = null) => throw new NotImplementedException();
-
-        public static Model NASNetMobile(TensorShape input_shape = null, bool include_top = true, string weights = "imagenet",
-                                    Tensor input_tensor = null, string pooling = null, int classes = 1000) => throw new NotImplementedException();
-
-        public static Model NASNetLarge(TensorShape input_shape = null, bool include_top = true, string weights = "imagenet",
-                                    Tensor input_tensor = null, string pooling = null, int classes = 1000) => throw new NotImplementedException();
-
-        public static Tensor _separable_conv_block(Tensor ip, int filters, (int, int)? kernel_size= null, (int, int)? strides= null, string block_id= null) => throw new NotImplementedException();
-
-        public static Tensor _adjust_block(Tensor p, Tensor ip, int filters, string block_id= null) => throw new NotImplementedException();
-
-        public static Tensor _normal_a_cell(Tensor p, Tensor ip, int filters, string block_id = null) => throw new NotImplementedException();
-
-        public static Tensor _reduction_a_cell(Tensor p, Tensor ip, int filters, string block_id = null) => throw new NotImplementedException();
-
-        public static Tensor preprocess_input(Tensor x, string data_format = null) => throw new NotImplementedException();
-
-        public static Tensor decode_predictions(Tensor preds, int top = 5) => throw new NotImplementedException();
-    }
-}

src/TensorFlowNET.Keras/Applications/Resnet.cs

@@ -1,41 +0,0 @@
-using System;
-using System.Collections.Generic;
-using System.Text;
-
-namespace Tensorflow.Keras.Applications
-{
-    public class Resnet
-    {
-        public static Model ResNet(Func<Tensor, Tensor> stack_fn, bool preact, bool use_bias, string model_name= "resnet", bool include_top= true,
-                                    string weights= "imagenet", Tensor input_tensor= null, TensorShape input_shape= null, string pooling= null,
-                                    int classes= 1000) => throw new NotImplementedException();
-
-        public static Tensor block1(Tensor x, int filters, int kernel_size= 3, int stride= 1, bool conv_shortcut= true, string name= null) => throw new NotImplementedException();
-
-        public static Tensor stack1(Tensor x, int filters, int blocks, int stride1 = 2, string name = null) => throw new NotImplementedException();
-
-        public static Tensor block2(Tensor x, int filters, int kernel_size = 3, int stride = 1, bool conv_shortcut = true, string name = null) => throw new NotImplementedException();
-
-        public static Tensor stack2(Tensor x, int filters, int blocks, int stride1 = 2, string name = null) => throw new NotImplementedException();
-
-        public static Tensor block3(Tensor x, int filters, int kernel_size = 3, int stride = 1, int groups = 32, bool conv_shortcut = true, string name = null) => throw new NotImplementedException();
-
-        public static Tensor stack3(Tensor x, int filters, int blocks, int stride1 = 2, int groups = 32, string name = null) => throw new NotImplementedException();
-
-        public static Model ResNet50(bool include_top = true, string weights = "imagenet",
-                                  Tensor input_tensor = null, TensorShape input_shape = null,
-                                  string pooling = null, int classes = 1000) => throw new NotImplementedException();
-
-        public static Model ResNet101(bool include_top = true, string weights = "imagenet",
-                                  Tensor input_tensor = null, TensorShape input_shape = null,
-                                  string pooling = null, int classes = 1000) => throw new NotImplementedException();
-
-        public static Model ResNet152(bool include_top = true, string weights = "imagenet",
-                                  Tensor input_tensor = null, TensorShape input_shape = null,
-                                  string pooling = null, int classes = 1000) => throw new NotImplementedException();
-
-        public static Tensor preprocess_input(Tensor x, string data_format = null) => throw new NotImplementedException();
-
-        public static Tensor decode_predictions(Tensor preds, int top = 5) => throw new NotImplementedException();
-    }
-}

src/TensorFlowNET.Keras/Applications/ResnetV2.cs

@@ -1,25 +0,0 @@
-using System;
-using System.Collections.Generic;
-using System.Text;
-
-namespace Tensorflow.Keras.Applications
-{
-    public class ResnetV2
-    {
-        public static Model ResNet50V2(bool include_top = true, string weights = "imagenet",
-                                   Tensor input_tensor = null, TensorShape input_shape = null,
-                                   string pooling = null, int classes = 1000) => throw new NotImplementedException();
-
-        public static Model ResNet101V2(bool include_top = true, string weights = "imagenet",
-                                  Tensor input_tensor = null, TensorShape input_shape = null,
-                                  string pooling = null, int classes = 1000) => throw new NotImplementedException();
-
-        public static Model ResNet152V2(bool include_top = true, string weights = "imagenet",
-                                  Tensor input_tensor = null, TensorShape input_shape = null,
-                                  string pooling = null, int classes = 1000) => throw new NotImplementedException();
-
-        public static Tensor preprocess_input(Tensor x, string data_format = null) => throw new NotImplementedException();
-
-        public static Tensor decode_predictions(Tensor preds, int top = 5) => throw new NotImplementedException();
-    }
-}

src/TensorFlowNET.Keras/Applications/Vgg16.cs

@@ -1,17 +0,0 @@
-using System;
-using System.Collections.Generic;
-using System.Text;
-
-namespace Tensorflow.Keras.Applications
-{
-    public class Vgg16
-    {
-        public static Model VGG16(bool include_top = true, string weights = "imagenet",
-                                Tensor input_tensor = null, TensorShape input_shape = null,
-                                string pooling = null, int classes = 1000) => throw new NotImplementedException();
-
-        public static Tensor preprocess_input(Tensor x, string data_format = null) => throw new NotImplementedException();
-
-        public static Tensor decode_predictions(Tensor preds, int top = 5) => throw new NotImplementedException();
-    }
-}

src/TensorFlowNET.Keras/Applications/Vgg19.cs

@@ -1,17 +0,0 @@
-using System;
-using System.Collections.Generic;
-using System.Text;
-
-namespace Tensorflow.Keras.Applications
-{
-    public class Vgg19
-    {
-        public static Model VGG19(bool include_top = true, string weights = "imagenet",
-                                Tensor input_tensor = null, TensorShape input_shape = null,
-                                string pooling = null, int classes = 1000) => throw new NotImplementedException();
-
-        public static Tensor preprocess_input(Tensor x, string data_format = null) => throw new NotImplementedException();
-
-        public static Tensor decode_predictions(Tensor preds, int top = 5) => throw new NotImplementedException();
-    }
-}

src/TensorFlowNET.Keras/Applications/Xception.cs

@@ -1,17 +0,0 @@
-using System;
-using System.Collections.Generic;
-using System.Text;
-
-namespace Tensorflow.Keras.Applications
-{
-    public class Xception
-    {
-        public static Model XCeption(bool include_top = true, string weights = "imagenet",
-                                Tensor input_tensor = null, TensorShape input_shape = null,
-                                string pooling = null, int classes = 1000) => throw new NotImplementedException();
-
-        public static Tensor preprocess_input(Tensor x, string data_format = null) => throw new NotImplementedException();
-
-        public static Tensor decode_predictions(Tensor preds, int top = 5) => throw new NotImplementedException();
-    }
-}

src/TensorFlowNET.Keras/Args.cs

@@ -1,29 +0,0 @@
-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/Backend.cs

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

src/TensorFlowNET.Keras/BackendConfig.cs

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

src/TensorFlowNET.Keras/Callbacks/BaseLogger.cs

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

src/TensorFlowNET.Keras/Callbacks/CSVLogger.cs

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

src/TensorFlowNET.Keras/Callbacks/Callback.cs

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

src/TensorFlowNET.Keras/Callbacks/CallbackList.cs

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

src/TensorFlowNET.Keras/Callbacks/EarlyStopping.cs

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

src/TensorFlowNET.Keras/Callbacks/History.cs

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

src/TensorFlowNET.Keras/Callbacks/LambdaCallback.cs

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

src/TensorFlowNET.Keras/Callbacks/LearningRateScheduler.cs

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

src/TensorFlowNET.Keras/Callbacks/ModelCheckpoint.cs

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

src/TensorFlowNET.Keras/Callbacks/ProgbarLogger.cs

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

src/TensorFlowNET.Keras/Callbacks/ReduceLROnPlateau.cs

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

src/TensorFlowNET.Keras/Callbacks/RemoteMonitor.cs

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

src/TensorFlowNET.Keras/Callbacks/TensorBoard.cs

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

src/TensorFlowNET.Keras/Callbacks/TensorBoardV1.cs

@@ -1,10 +0,0 @@
-using System;
-using System.Collections.Generic;
-using System.Text;
-
-namespace Tensorflow.Keras.Callbacks
-{
-    class TensorBoardV1
-    {
-    }
-}

src/TensorFlowNET.Keras/Callbacks/TerminateOnNaN.cs

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

src/TensorFlowNET.Keras/Constraints/ConstraintBase.cs

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

src/TensorFlowNET.Keras/Constraints/MaxNorm.cs

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

src/TensorFlowNET.Keras/Constraints/MinMaxNorm.cs

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

src/TensorFlowNET.Keras/Constraints/NonNeg.cs

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

src/TensorFlowNET.Keras/Constraints/RadialConstraint.cs

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

src/TensorFlowNET.Keras/Constraints/UnitNorm.cs

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

src/TensorFlowNET.Keras/Core.cs

@@ -1,13 +0,0 @@
-using Tensorflow;
-using static Tensorflow.Binding;
-
-namespace Keras
-{
-    public static class Keras
-    {
-        public static Tensor create_tensor(int[] shape, float mean = 0, float stddev = 1, TF_DataType dtype = TF_DataType.TF_FLOAT, int? seed = null, string name = null)
-        {
-            return tf.truncated_normal(shape: shape, mean: mean, stddev: stddev, dtype: dtype, seed: seed, name: name);
-        }
-    }
-}

src/TensorFlowNET.Keras/Datasets/BostonHousing.cs

@@ -1,11 +0,0 @@
-using System;
-using System.Collections.Generic;
-using System.Text;
-
-namespace Tensorflow.Keras.Datasets
-{
-    public class BostonHousing
-    {
-        public static ((Tensor, Tensor), (Tensor, Tensor)) load_data(string path = "boston_housing.npz", float test_split = 0.2f, int seed = 113) => throw new NotImplementedException();
-    }
-}

src/TensorFlowNET.Keras/Datasets/Cifar.cs

@@ -1,11 +0,0 @@
-using System;
-using System.Collections.Generic;
-using System.Text;
-
-namespace Tensorflow.Keras.Datasets
-{
-    public class Cifar
-    {
-        public (Tensor, Tensor) load_batch(string fpath, string label_key = "labels") => throw new NotImplementedException();
-    }
-}

src/TensorFlowNET.Keras/Datasets/Cifar10.cs

@@ -1,11 +0,0 @@
-using System;
-using System.Collections.Generic;
-using System.Text;
-
-namespace Tensorflow.Keras.Datasets
-{
-    public class Cifar10
-    {
-        public static ((Tensor, Tensor), (Tensor, Tensor)) load_data() => throw new NotImplementedException();
-    }
-}

src/TensorFlowNET.Keras/Datasets/Cifar100.cs

@@ -1,11 +0,0 @@
-using System;
-using System.Collections.Generic;
-using System.Text;
-
-namespace Tensorflow.Keras.Datasets
-{
-    public class Cifar100
-    {
-        public static ((Tensor, Tensor), (Tensor, Tensor)) load_data(string label_mode = "fine") => throw new NotImplementedException();
-    }
-}

src/TensorFlowNET.Keras/Datasets/FashionMNIST.cs

@@ -1,11 +0,0 @@
-using System;
-using System.Collections.Generic;
-using System.Text;
-
-namespace Tensorflow.Keras.Datasets
-{
-    public class FashionMNIST
-    {
-        public static ((Tensor, Tensor), (Tensor, Tensor)) load_data() => throw new NotImplementedException();
-    }
-}

src/TensorFlowNET.Keras/Datasets/IMDB.cs

@@ -1,15 +0,0 @@
-using Newtonsoft.Json.Linq;
-using System;
-using System.Collections.Generic;
-using System.Text;
-
-namespace Tensorflow.Keras.Datasets
-{
-    public class IMDB
-    {
-        public static ((Tensor, Tensor), (Tensor, Tensor)) load_data(string path= "imdb.npz", int? num_words= null, int skip_top= 0, int? maxlen= null,
-                                                            int seed= 113,int start_char= 1, int oov_char= 2, int index_from= 3) => throw new NotImplementedException();
-
-        public static JObject get_word_index(string path= "imdb_word_index.json") => throw new NotImplementedException();
-    }
-}

src/TensorFlowNET.Keras/Datasets/MNIST.cs

@@ -1,11 +1,74 @@
-using System;
-using System.Collections.Generic;
-using System.Text;
+/*****************************************************************************
+   Copyright 2020 Haiping Chen. All Rights Reserved.
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.
+******************************************************************************/
+
+using NumSharp;
+using System;
+using System.IO;
+using System.Net;
 
 namespace Tensorflow.Keras.Datasets
 {
-    public class MNIST
+    public class Mnist
     {
-        public static ((Tensor, Tensor), (Tensor, Tensor)) load_data(string path = "mnist.npz") => throw new NotImplementedException();
+        string origin_folder = "https://storage.googleapis.com/tensorflow/tf-keras-datasets/";
+        string file_name = "mnist.npz";
+
+        /// <summary>
+        /// Loads the [MNIST dataset](http://yann.lecun.com/exdb/mnist/).
+        /// </summary>
+        /// <returns></returns>
+        public DatasetPass load_data()
+        {
+            var file = Download();
+            var bytes = File.ReadAllBytes(file);
+            var datax = LoadX(bytes);
+            var datay = LoadY(bytes);
+            return new DatasetPass
+            {
+                Train = (datax.Item1, datay.Item1),
+                Test = (datax.Item2, datay.Item2)
+            };
+        }
+
+        (NDArray, NDArray) LoadX(byte[] bytes)
+        {
+            var y = np.Load_Npz<byte[,,]>(bytes);
+            return (y["x_train.npy"], y["x_test.npy"]);
+        }
+
+        (NDArray, NDArray) LoadY(byte[] bytes)
+        {
+            var y = np.Load_Npz<byte[]>(bytes);
+            return (y["y_train.npy"], y["y_test.npy"]);
+        }
+
+        string Download()
+        {
+            var fileSaveTo = Path.Combine(Path.GetTempPath(), file_name);
+
+            if (File.Exists(fileSaveTo))
+            {
+                Console.WriteLine($"The file {fileSaveTo} already exists");
+                return fileSaveTo;
+            }
+
+            using var wc = new WebClient();
+            wc.DownloadFileTaskAsync(origin_folder + file_name, fileSaveTo).Wait();
+
+            return fileSaveTo;
+        }
     }
 }

src/TensorFlowNET.Keras/Datasets/Reuters.cs

@@ -1,12 +0,0 @@
-using System;
-using System.Collections.Generic;
-using System.Text;
-
-namespace Tensorflow.Keras.Datasets
-{
-    public class Reuters
-    {
-        public static ((Tensor, Tensor), (Tensor, Tensor)) load_data(string path = "reuters.npz", int? num_words= null, int skip_top= 0,
-                                                int? maxlen= null,float test_split= 0.2f, int seed= 113,int start_char= 1,int oov_char= 2,int index_from= 3) => throw new NotImplementedException();
-    }
-}

src/TensorFlowNET.Keras/Distribute/DistributedTrainingUtils.cs

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

src/TensorFlowNET.Keras/Distribute/KerasCorrectnessTestBase.cs

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

src/TensorFlowNET.Keras/Distribute/KerasDnnCorrectnessTest.cs

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

src/TensorFlowNET.Keras/Distribute/KerasEmbeddingModelCorrectnessTest.cs

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

src/TensorFlowNET.Keras/Distribute/KerasImageModelCorrectnessTest.cs

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

src/TensorFlowNET.Keras/Distribute/KerasOptimizerV2Test.cs

@@ -1,10 +0,0 @@
-using System;
-using System.Collections.Generic;
-using System.Text;
-
-namespace Tensorflow.Keras.Distribute
-{
-    class KerasOptimizerV2Test
-    {
-    }
-}

src/TensorFlowNET.Keras/Distribute/KerasPremadeModelsTest.cs

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

src/TensorFlowNET.Keras/Distribute/KerasRnnModelCorrectnessTest.cs

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

src/TensorFlowNET.Keras/Distribute/KerasStatefulLstmModelCorrectnessTest.cs

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

src/TensorFlowNET.Keras/Distribute/KerasUtilsTest.cs

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