Fix Sequential model.summary missing layers. SciSharp#960

Oceania2018 · Oceania2018 · commit ea6826a7cb5b · 2022-12-03T12:44:00.000-06:00
diff --git a/SciSharp STACK Examples.sln b/SciSharp STACK Examples.sln
@@ -23,6 +23,8 @@ Project("{9A19103F-16F7-4668-BE54-9A1E7A4F7556}") = "SharpCV", "..\SharpCV\src\S
 EndProject
 Project("{9A19103F-16F7-4668-BE54-9A1E7A4F7556}") = "SciSharp.Models.ImageClassification", "..\SciSharp.Models\SciSharp.Models.ImageClassification\SciSharp.Models.ImageClassification.csproj", "{FA0B3C1B-9364-40D9-A7E8-54A6045F7B67}"
 EndProject
+Project("{9A19103F-16F7-4668-BE54-9A1E7A4F7556}") = "SciSharp.Models.TextClassification", "..\SciSharp.Models\SciSharp.Models.TextClassification\SciSharp.Models.TextClassification.csproj", "{D5243F2C-4EAE-45AD-849C-A7FD79B599C5}"
+EndProject
 Global
 	GlobalSection(SolutionConfigurationPlatforms) = preSolution
 		Debug|Any CPU = Debug|Any CPU
@@ -216,6 +218,24 @@ Global
 		{FA0B3C1B-9364-40D9-A7E8-54A6045F7B67}.Release|x64.Build.0 = Release|Any CPU
 		{FA0B3C1B-9364-40D9-A7E8-54A6045F7B67}.Release|x86.ActiveCfg = Release|Any CPU
 		{FA0B3C1B-9364-40D9-A7E8-54A6045F7B67}.Release|x86.Build.0 = Release|Any CPU
+		{D5243F2C-4EAE-45AD-849C-A7FD79B599C5}.Debug|Any CPU.ActiveCfg = Debug|Any CPU
+		{D5243F2C-4EAE-45AD-849C-A7FD79B599C5}.Debug|Any CPU.Build.0 = Debug|Any CPU
+		{D5243F2C-4EAE-45AD-849C-A7FD79B599C5}.Debug|x64.ActiveCfg = Debug|x64
+		{D5243F2C-4EAE-45AD-849C-A7FD79B599C5}.Debug|x64.Build.0 = Debug|x64
+		{D5243F2C-4EAE-45AD-849C-A7FD79B599C5}.Debug|x86.ActiveCfg = Debug|Any CPU
+		{D5243F2C-4EAE-45AD-849C-A7FD79B599C5}.Debug|x86.Build.0 = Debug|Any CPU
+		{D5243F2C-4EAE-45AD-849C-A7FD79B599C5}.GPU|Any CPU.ActiveCfg = Debug|Any CPU
+		{D5243F2C-4EAE-45AD-849C-A7FD79B599C5}.GPU|Any CPU.Build.0 = Debug|Any CPU
+		{D5243F2C-4EAE-45AD-849C-A7FD79B599C5}.GPU|x64.ActiveCfg = Debug|x64
+		{D5243F2C-4EAE-45AD-849C-A7FD79B599C5}.GPU|x64.Build.0 = Debug|x64
+		{D5243F2C-4EAE-45AD-849C-A7FD79B599C5}.GPU|x86.ActiveCfg = Debug|Any CPU
+		{D5243F2C-4EAE-45AD-849C-A7FD79B599C5}.GPU|x86.Build.0 = Debug|Any CPU
+		{D5243F2C-4EAE-45AD-849C-A7FD79B599C5}.Release|Any CPU.ActiveCfg = Release|Any CPU
+		{D5243F2C-4EAE-45AD-849C-A7FD79B599C5}.Release|Any CPU.Build.0 = Release|Any CPU
+		{D5243F2C-4EAE-45AD-849C-A7FD79B599C5}.Release|x64.ActiveCfg = Release|x64
+		{D5243F2C-4EAE-45AD-849C-A7FD79B599C5}.Release|x64.Build.0 = Release|x64
+		{D5243F2C-4EAE-45AD-849C-A7FD79B599C5}.Release|x86.ActiveCfg = Release|Any CPU
+		{D5243F2C-4EAE-45AD-849C-A7FD79B599C5}.Release|x86.Build.0 = Release|Any CPU
 	EndGlobalSection
 	GlobalSection(SolutionProperties) = preSolution
 		HideSolutionNode = FALSE
diff --git a/src/TensorFlowNET.Examples/GAN/MnistGAN.cs b/src/TensorFlowNET.Examples/GAN/MnistGAN.cs
@@ -16,8 +16,12 @@ namespace TensorFlowNET.Examples.GAN
     public class MnistGAN : SciSharpExample, IExample
     {
         float LeakyReLU_alpha = 0.2f;
+
+#if GPU
+        int epochs = 2000; // Better effect, but longer time
+#else
         int epochs = 20;
-        //int epochs = 2000; // Better effect, but longer time
+#endif
         int batch_size = 64;
 
         string imgpath = "dcgan\\imgs";
@@ -65,7 +69,8 @@ private Model Make_Generator_model()
             Tensorflow.Keras.Activation activation = null;
 
             var model = keras.Sequential();
-            model.add(keras.layers.Dense(img_rows / 4 * img_cols / 4 * 256, activation: activation, input_shape: 100));
+            model.add(keras.layers.Input(shape: 100));
+            model.add(keras.layers.Dense(img_rows / 4 * img_cols / 4 * 256, activation: activation));
             model.add(keras.layers.BatchNormalization(momentum: 0.8f));
             model.add(keras.layers.LeakyReLU(LeakyReLU_alpha));
             model.add(keras.layers.Reshape((7, 7, 256)));
diff --git a/src/TensorFlowNET.Examples/NaturalLanguageProcessing/SentimentClassification.cs b/src/TensorFlowNET.Examples/NaturalLanguageProcessing/SentimentClassification.cs
@@ -0,0 +1,103 @@
+﻿using SciSharp.Models.TimeSeries;
+using SciSharp.Models;
+using System;
+using System.Collections.Generic;
+using System.Linq;
+using System.Text;
+using System.Threading.Tasks;
+using System.IO;
+using static Tensorflow.Binding;
+using static Tensorflow.KerasApi;
+
+namespace TensorFlowNET.Examples
+{
+    public class SentimentClassification : SciSharpExample, IExample
+    {
+        ITimeSeriesTask task;
+        public ExampleConfig InitConfig()
+            => Config = new ExampleConfig
+            {
+                Name = "Text Sentiment Classification",
+                Enabled = true
+            };
+
+        public bool Run()
+        {
+            var wizard = new ModelWizard();
+            task = wizard.AddTimeSeriesTask<ConvolutionalModel>(new TaskOptions
+            {
+                WeightsPath = @"timeseries_linear_v1\saved_weights.h5"
+            });
+            task.SetModelArgs(new TimeSeriesModelArgs
+            {
+            });
+
+            return true;
+        }
+
+        public override void PrepareData()
+        {
+            // tf.debugging.set_log_device_placement(true);
+            string url = "https://ai.stanford.edu/~amaas/data/sentiment/aclImdb_v1.tar.gz";
+            var dataset = keras.utils.get_file("aclImdb_v1.tar.gz", url,
+                untar: true,
+                cache_dir: Path.GetTempPath(),
+                cache_subdir: "aclImdb_v1");
+            var data_dir = Path.Combine(dataset, "aclImdb");
+            var train_dir = Path.Combine(data_dir, "train");
+
+            int batch_size = 32;
+            int seed = 42;
+            var raw_train_ds = keras.preprocessing.text_dataset_from_directory(
+                train_dir,
+                batch_size: batch_size,
+                validation_split: 0.2f,
+                subset: "training",
+                seed: seed);
+
+            /*foreach (var (text_batch, label_batch) in raw_train_ds.take(1))
+            {
+                foreach (var i in range(3))
+                {
+                    print("Review", text_batch.StringData()[i]);
+                    print("Label", label_batch.numpy()[i]);
+                }
+            }
+
+            print("Label 0 corresponds to", raw_train_ds.class_names[0]);
+            print("Label 1 corresponds to", raw_train_ds.class_names[1]);
+
+            var raw_val_ds = keras.preprocessing.text_dataset_from_directory(
+                train_dir,
+                batch_size: batch_size,
+                validation_split: 0.2f,
+                subset: "validation",
+            seed: seed);
+
+            var test_dir = Path.Combine(data_dir, "test");
+            var raw_test_ds = keras.preprocessing.text_dataset_from_directory(
+                test_dir,
+                batch_size: batch_size);
+
+            var max_features = 10000;
+            var sequence_length = 250;
+
+            Func<Tensor, Tensor> custom_standardization = input_data =>
+            {
+                var lowercase = tf.strings.lower(input_data);
+                var stripped_html = tf.strings.regex_replace(lowercase, "<br />", " ");
+                return tf.strings.regex_replace(stripped_html,
+                                                "'[!\"\\#\\$%\\&\'\\(\\)\\*\\+,\\-\\./:;<=>\\?@\\[\\\\\\]\\^_`\\{\\|\\}\\~]'",
+                "");
+            };
+
+            var vectorize_layer = keras.layers.preprocessing.TextVectorization(standardize: custom_standardization,
+                max_tokens: max_features,
+                output_mode: "int",
+                output_sequence_length: sequence_length);
+
+            var train_text = raw_train_ds.map(inputs => inputs[0]);
+            //vectorize_layer.adapt(train_text);*/
+        }
+    }
+}
diff --git a/src/TensorFlowNET.Examples/ObjectDetection/MnistInYOLOv3.cs b/src/TensorFlowNET.Examples/ObjectDetection/MnistInYOLOv3.cs
@@ -31,14 +31,15 @@ public ExampleConfig InitConfig()
             => Config = new ExampleConfig
             {
                 Name = "MNIST in YOLOv3",
-                Enabled = true
+                Enabled = false
             };
 
         public bool Run()
         {
             cfg = new YoloConfig("YOLOv3");
             (trainingData, testingData) = PrepareData();
             Train();
+            Test();
             return true;
         }
 
@@ -67,6 +68,7 @@ public override void Test()
             {
                 ModelPath = @"./YOLOv3/yolov3.h5"
             });
+            task.SetModelArgs(cfg);
             var result = task.Test(new TestingOptions
             {
                 
diff --git a/src/TensorFlowNET.Examples/Program.cs b/src/TensorFlowNET.Examples/Program.cs
@@ -36,8 +36,8 @@ static void Main(string[] args)
             var examples = Assembly.GetEntryAssembly().GetTypes()
                 .Where(x => x.GetInterfaces().Contains(typeof(IExample)))
                 //.Where(x => x.Name == nameof(WeatherPrediction))
-                //.Where(x => x.Name == nameof(NeuralNetXorEager))
-                .Where(x => x.Name == nameof(MnistInYOLOv3))
+                //.Where(x => x.Name == nameof(SentimentClassification))
+                //.Where(x => x.Name == nameof(MnistInYOLOv3))
                 .ToArray();
 
             Console.WriteLine(Environment.OSVersion, Color.Yellow);
diff --git a/src/TensorFlowNET.Examples/TensorFlowNET.Examples.csproj b/src/TensorFlowNET.Examples/TensorFlowNET.Examples.csproj
@@ -61,6 +61,7 @@
   <ItemGroup>
     <ProjectReference Include="..\..\..\SciSharp.Models\SciSharp.Models.ImageClassification\SciSharp.Models.ImageClassification.csproj" />
     <ProjectReference Include="..\..\..\SciSharp.Models\SciSharp.Models.ObjectDetection\SciSharp.Models.ObjectDetection.csproj" />
+    <ProjectReference Include="..\..\..\SciSharp.Models\SciSharp.Models.TextClassification\SciSharp.Models.TextClassification.csproj" />
     <ProjectReference Include="..\..\..\SciSharp.Models\SciSharp.Models.TimeSeries\SciSharp.Models.TimeSeries.csproj" />
     <ProjectReference Include="..\..\..\SharpCV\src\SharpCV\SharpCV.csproj" />
     <ProjectReference Include="..\..\..\TensorFlow.NET\src\TensorFlowNET.Core\Tensorflow.Binding.csproj" />

Original file line number	Diff line number	Diff line change
`@@ -31,14 +31,15 @@ public ExampleConfig InitConfig()`
`31`	`31`	`=> Config = new ExampleConfig`
`32`	`32`	`{`
`33`	`33`	`Name = "MNIST in YOLOv3",`
`34`		`- Enabled = true`
	`34`	`+ Enabled = false`
`35`	`35`	`};`
`36`	`36`
`37`	`37`	`public bool Run()`
`38`	`38`	`{`
`39`	`39`	`cfg = new YoloConfig("YOLOv3");`
`40`	`40`	`(trainingData, testingData) = PrepareData();`
`41`	`41`	`Train();`
	`42`	`+ Test();`
`42`	`43`	`return true;`
`43`	`44`	`}`
`44`	`45`
`@@ -67,6 +68,7 @@ public override void Test()`
`67`	`68`	`{`
`68`	`69`	`ModelPath = @"./YOLOv3/yolov3.h5"`
`69`	`70`	`});`
	`71`	`+ task.SetModelArgs(cfg);`
`70`	`72`	`var result = task.Test(new TestingOptions`
`71`	`73`	`{`
`72`	`74`