Fixed tutorials and examples

Kotlin · May 14, 2021 · dc5883a · dc5883a
1 parent 7b7c48d
commit dc5883a
Show file tree

Hide file tree

Showing 6 changed files with 25 additions and 24 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -1,4 +1,4 @@
-# 0.2.0 (01/05/2021)
+# 0.2.0 (17/05/2021)
 Features:
 * Added [support for Functional API](https://github.com/JetBrains/KotlinDL/issues/23)
 * Added [BatchNorm layer](https://github.com/JetBrains/KotlinDL/issues/34) for inference

diff --git a/README.md b/README.md
@@ -148,8 +148,8 @@ To use KotlinDL in your project, add the following dependency to your `build.gra
        implementation 'org.jetbrains.kotlinx:kotlin-deeplearning-api:[KOTLIN-DL-VERSION]'
    }
 ```
-The latest KotlinDL version is 0.2.0-alpha-1. 
-The latest stable KotlinDL version is 0.1.1. 
+The latest KotlinDL version is 0.2.0. 
+The latest stable KotlinDL version is 0.2.0. 
 
 For more details, as well as for `pom.xml` and `build.gradle.kts` examples, please refer to the [Quick Start Guide](docs/quick_start_guide.md).
 

diff --git a/build.gradle b/build.gradle
@@ -8,7 +8,7 @@ apply plugin: "io.codearte.nexus-staging"
 
 allprojects {
     group = 'org.jetbrains.kotlinx'
-    version = '0.2.0-alpha-2'
+    version = '0.2.0'
 
     apply plugin: "maven"
     apply plugin: "kotlin"

diff --git a/docs/transfer_learning.md b/docs/transfer_learning.md
@@ -27,24 +27,26 @@ You can do so via the Image Preprocessing Pipeline description, and building a d
 Here's code that will go through a folder structure received via ```catDogsSmallDatasetPath()```, loads and resizes the images, and applies the VGG-19 specific preprocessing.
 
 ```kotlin
-val catdogimages = catDogsSmallDatasetPath()
+val dogsVsCatsDatasetPath = catDogsSmallDatasetPath()
 
-val preprocessing: Preprocessing = preprocessingPipeline {
-    imagePreprocessing {
-         load {
-            pathToData = File(catdogimages)
-            imageShape = ImageShape(channels = 3)
+val preprocessing: Preprocessing = preprocess {
+    transformImage {
+        load {
+            pathToData = File(dogsVsCatsDatasetPath)
+            imageShape = ImageShape(channels = NUM_CHANNELS)
             colorMode = ColorOrder.BGR
             labelGenerator = FromFolders(mapping = mapOf("cat" to 0, "dog" to 1))
-         }
-         resize {
-            outputHeight = 224
-            outputWidth = 224
+        }
+        resize {
+            outputHeight = IMAGE_SIZE.toInt()
+            outputWidth = IMAGE_SIZE.toInt()
             interpolation = InterpolationType.BILINEAR
-         }
+        }
     }
-    sharpen {
-        modelType = ModelType.VGG_19
+    transformTensor {
+        sharpen {
+            modelType = ModelType.VGG_19
+        }
     }
 }
 

diff --git a/examples/src/main/kotlin/examples/dataset/ImagePreprocessing.kt b/examples/src/main/kotlin/examples/dataset/ImagePreprocessing.kt
@@ -51,7 +51,7 @@ fun main() {
                 colorMode = ColorOrder.BGR
             }
             rotate {
-                degrees = 30f
+                degrees = 60f
             }
             crop {
                 left = 12
@@ -60,8 +60,8 @@ fun main() {
                 bottom = 12
             }
             resize {
-                outputWidth = 400
-                outputHeight = 400
+                outputWidth = 300
+                outputHeight = 300
                 interpolation = InterpolationType.NEAREST
             }
         }
@@ -81,7 +81,7 @@ fun main() {
     val rawImage = batchIter.next().x[1]
 
     val frame = JFrame("Filters")
-    frame.contentPane.add(ImagesJPanel(rawImage, ImageShape(400, 400, 3)))
+    frame.contentPane.add(ImagesJPanel(rawImage, ImageShape(300, 300, 3)))
     frame.setSize(1000, 1000)
     frame.isVisible = true
     frame.defaultCloseOperation = JFrame.EXIT_ON_CLOSE

diff --git a/...in/kotlin/examples/transferlearning/modelzoo/vgg19/Example_3_VGG19_additional_training.kt b/...in/kotlin/examples/transferlearning/modelzoo/vgg19/Example_3_VGG19_additional_training.kt
@@ -56,12 +56,12 @@ fun vgg19additionalTraining() {
     val modelZoo = ModelZoo(commonModelDirectory = File("cache/pretrainedModels"), modelType = ModelType.VGG_19)
     val model = modelZoo.loadModel() as Sequential
 
-    val catdogimages = dogsCatsSmallDatasetPath()
+    val dogsVsCatsDatasetPath = dogsCatsSmallDatasetPath()
 
     val preprocessing: Preprocessing = preprocess {
         transformImage {
             load {
-                pathToData = File(catdogimages)
+                pathToData = File(dogsVsCatsDatasetPath)
                 imageShape = ImageShape(channels = NUM_CHANNELS)
                 colorMode = ColorOrder.BGR
                 labelGenerator = FromFolders(mapping = mapOf("cat" to 0, "dog" to 1))
@@ -77,7 +77,6 @@ fun vgg19additionalTraining() {
                 modelType = ModelType.VGG_19
             }
         }
-
     }
 
     val dataset = OnFlyImageDataset.create(preprocessing).shuffle()