Convert MNIST to Epochs

Autoencoder/Autoencoder1D/main.swift

@@ -24,7 +24,7 @@ let imageHeight = 28
 let imageWidth = 28
 
 let outputFolder = "./output/"
-let dataset = FashionMNIST(batchSize: batchSize, flattening: true)
+let dataset = OldFashionMNIST(batchSize: batchSize, flattening: true)
 // An autoencoder.
 var autoencoder = Sequential {
     // The encoder.

Autoencoder/Autoencoder2D/main.swift

@@ -26,7 +26,7 @@ let imageHeight = 28
 let imageWidth = 28
 
 let outputFolder = "./output/"
-let dataset = KuzushijiMNIST(batchSize: batchSize, flattening: true)
+let dataset = OldKuzushijiMNIST(batchSize: batchSize, flattening: true)
 
 // An autoencoder.
 struct Autoencoder2D: Layer {

Autoencoder/VAE1D/main.swift

@@ -26,7 +26,7 @@ let imageHeight = 28
 let imageWidth = 28
 
 let outputFolder = "./output/"
-let dataset = MNIST(batchSize: 128, flattening: true)
+let dataset = OldMNIST(batchSize: 128, flattening: true)
 
 let inputDim = 784  // 28*28 for any MNIST
 let hiddenDim = 400
@@ -84,7 +84,7 @@ func vaeLossFunction(
 }
 
 // TODO: Find a cleaner way of extracting individual images that doesn't require a second dataset.
-let singleImageDataset = MNIST(batchSize: 1, flattening: true)
+let singleImageDataset = OldMNIST(batchSize: 1, flattening: true)
 let individualTestImages = singleImageDataset.test
 var testImageIterator = individualTestImages.sequenced()
 

Benchmarks/Models/LeNetMnist.swift

@@ -37,11 +37,11 @@ enum LeNetMNIST: BenchmarkModel {
     }
 
     static func makeInferenceBenchmark(settings: BenchmarkSettings) -> Benchmark {
-        return ImageClassificationInference<LeNet, MNIST>(settings: settings)
+        return ImageClassificationInference<LeNet, OldMNIST>(settings: settings)
     }
 
     static func makeTrainingBenchmark(settings: BenchmarkSettings) -> Benchmark {
-        return ImageClassificationTraining<LeNet, MNIST>(settings: settings)
+        return ImageClassificationTraining<LeNet, OldMNIST>(settings: settings)
     }
 }
 

DCGAN/main.swift

@@ -18,7 +18,7 @@ import ModelSupport
 import TensorFlow
 
 let batchSize = 512
-let mnist = MNIST(batchSize: batchSize, flattening: false, normalizing: true)
+let mnist = OldMNIST(batchSize: batchSize, flattening: false, normalizing: true)
 
 let outputFolder = "./output/"
 

Datasets/CIFAR10/CIFAR10.swift

@@ -48,12 +48,13 @@ public struct CIFAR10<Entropy: RandomNumberGenerator> {
     self.init(
       batchSize: batchSize,
       entropy: entropy,
+      device: Device.default,
       remoteBinaryArchiveLocation: URL(
         string: "https://storage.googleapis.com/s4tf-hosted-binaries/datasets/CIFAR10/cifar-10-binary.tar.gz")!, 
       normalizing: true)
   }
 
-  /// Creates an instance with `batchSize` using `remoteBinaryArchiveLocation`.
+  /// Creates an instance with `batchSize` on `device` using `remoteBinaryArchiveLocation`.
   ///
   /// - Parameters:
   ///   - entropy: a source of randomness used to shuffle sample ordering.  It  
@@ -65,6 +66,7 @@ public struct CIFAR10<Entropy: RandomNumberGenerator> {
   public init(
     batchSize: Int,
     entropy: Entropy,
+    device: Device,
     remoteBinaryArchiveLocation: URL, 
     localStorageDirectory: URL = DatasetUtilities.defaultDirectory
       .appendingPathComponent("CIFAR10", isDirectory: true), 
@@ -76,13 +78,13 @@ public struct CIFAR10<Entropy: RandomNumberGenerator> {
     let trainingSamples = loadCIFARTrainingFiles(in: localStorageDirectory)
     training = TrainingEpochs(samples: trainingSamples, batchSize: batchSize, entropy: entropy)
       .lazy.map { (batches: Batches) -> LazyMapSequence<Batches, LabeledImage> in
-        return batches.lazy.map{ makeBatch(samples: $0, normalizing: normalizing) }
+        return batches.lazy.map{ makeBatch(samples: $0, normalizing: normalizing, device:device) }
       }
 
     // Validation data
     let validationSamples = loadCIFARTestFile(in: localStorageDirectory)
     validation = validationSamples.inBatches(of: batchSize).lazy.map {
-      makeBatch(samples: $0, normalizing: normalizing)
+      makeBatch(samples: $0, normalizing: normalizing, device:device)
     }
   }
 }
@@ -145,16 +147,17 @@ func loadCIFARTestFile(in localStorageDirectory: URL) -> [(data: [UInt8], label:
   return loadCIFARFile(named: "test_batch.bin", in: localStorageDirectory)
 }
 
-func makeBatch<BatchSamples: Collection>(samples: BatchSamples, normalizing: Bool) -> LabeledImage 
-where BatchSamples.Element == (data: [UInt8], label: Int32) {
+fileprivate func makeBatch<BatchSamples: Collection>(
+  samples: BatchSamples, normalizing: Bool, device:Device
+) -> LabeledImage where BatchSamples.Element == (data: [UInt8], label: Int32) {
   let bytes = samples.lazy.map(\.data).reduce(into: [], +=)
-  let images = Tensor<UInt8>(shape: [samples.count, 3, 32, 32], scalars: bytes)
+  let images = Tensor<UInt8>(shape: [samples.count, 3, 32, 32], scalars: bytes, on:device)
 
   var imageTensor = Tensor<Float>(images.transposed(permutation: [0, 2, 3, 1]))
   imageTensor /= 255.0
   if normalizing {
-    let mean = Tensor<Float>([0.4913996898, 0.4821584196, 0.4465309242])
-    let std = Tensor<Float>([0.2470322324, 0.2434851280, 0.2615878417])
+    let mean = Tensor<Float>([0.4913996898, 0.4821584196, 0.4465309242], on:device)
+    let std = Tensor<Float>([0.2470322324, 0.2434851280, 0.2615878417], on:device)
     imageTensor = (imageTensor - mean) / std
   }
 

Datasets/CMakeLists.txt

@@ -15,6 +15,10 @@ add_library(Datasets
   MNIST/MNIST.swift
   MNIST/FashionMNIST.swift
   MNIST/KuzushijiMNIST.swift
+  MNIST/OldMNISTDatasetHandler.swift
+  MNIST/OldMNIST.swift
+  MNIST/OldFashionMNIST.swift
+  MNIST/OldKuzushijiMNIST.swift
   ObjectDetectionDataset.swift
   BostonHousing/BostonHousing.swift
   TextUnsupervised/TextUnsupervised.swift

Datasets/MNIST/FashionMNIST.swift

@@ -21,41 +21,102 @@ import Foundation
 import TensorFlow
 import Batcher
 
-public struct FashionMNIST: ImageClassificationDataset {
-    public typealias SourceDataSet = [TensorPair<Float, Int32>]
-    public let training: Batcher<SourceDataSet>
-    public let test: Batcher<SourceDataSet>
+// Copyright 2019 The TensorFlow 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.
 
-    public init(batchSize: Int) {
-        self.init(batchSize: batchSize, flattening: false, normalizing: false)
-    }
+// Original source:
+// "The MNIST database of handwritten digits"
+// Yann LeCun, Corinna Cortes, and Christopher J.C. Burges
+// http://yann.lecun.com/exdb/mnist/
 
-    public init(
-        batchSize: Int, flattening: Bool = false, normalizing: Bool = false,
-        localStorageDirectory: URL = DatasetUtilities.defaultDirectory
-            .appendingPathComponent("FashionMNIST", isDirectory: true)
-    ) {
-        training = Batcher<SourceDataSet>(
-            on: fetchMNISTDataset(
-                localStorageDirectory: localStorageDirectory,
-                remoteBaseDirectory: "http://fashion-mnist.s3-website.eu-central-1.amazonaws.com/",
-                imagesFilename: "train-images-idx3-ubyte",
-                labelsFilename: "train-labels-idx1-ubyte",
-                flattening: flattening,
-                normalizing: normalizing),
-            batchSize: batchSize,
-            numWorkers: 1, //No need to use parallelism since everything is loaded in memory
-            shuffle: true)
-
-        test = Batcher<SourceDataSet>(
-            on: fetchMNISTDataset(
-                localStorageDirectory: localStorageDirectory,
-                remoteBaseDirectory: "http://fashion-mnist.s3-website.eu-central-1.amazonaws.com/",
-                imagesFilename: "t10k-images-idx3-ubyte",
-                labelsFilename: "t10k-labels-idx1-ubyte",
-                flattening: flattening,
-                normalizing: normalizing),
-            batchSize: batchSize,
-            numWorkers: 1) //No need to use parallelism since everything is loaded in memory
+import Foundation
+import TensorFlow
+import Batcher
+
+public struct FashionMNIST<Entropy: RandomNumberGenerator> {
+  /// Type of the collection of non-collated batches.
+  public typealias Batches = Slices<Sampling<[(data: [UInt8], label: Int32)], ArraySlice<Int>>>
+  /// The type of the training data, represented as a sequence of epochs, which
+  /// are collection of batches.
+  public typealias Training = LazyMapSequence<
+    TrainingEpochs<[(data: [UInt8], label: Int32)], Entropy>,
+    LazyMapSequence<Batches, LabeledImage>
+  >
+  /// The type of the validation data, represented as a collection of batches.
+  public typealias Validation = LazyMapSequence<Slices<[(data: [UInt8], label: Int32)]>, LabeledImage>
+  /// The training epochs.
+  public let training: Training
+  /// The validation batches.
+  public let validation: Validation
+
+  /// Creates an instance with `batchSize`.
+  ///
+  /// - Parameter entropy: a source of randomness used to shuffle sample 
+  ///   ordering.  It  will be stored in `self`, so if it is only pseudorandom 
+  ///   and has value semantics, the sequence of epochs is deterministic and not 
+  ///   dependent on other operations.
+  public init(batchSize: Int, entropy: Entropy) {
+    self.init(batchSize: batchSize, device: Device.default, entropy: entropy,
+              flattening: false, normalizing: false)
+  }
+
+  /// Creates an instance with `batchSize` on `device`.
+  ///
+  /// - Parameters:
+  ///   - entropy: a source of randomness used to shuffle sample ordering.  It  
+  ///     will be stored in `self`, so if it is only pseudorandom and has value 
+  ///     semantics, the sequence of epochs is deterministic and not dependent 
+  ///     on other operations.
+  ///   - flattening: flattens the data to be a 2d-tensor iff `true. The default value
+  ///     is `false`.
+  ///   - normalizing: normalizes the batches to have values from -1.0 to 1.0 iff `true`.
+  ///     The default value is `false`.
+  ///   - localStorageDirectory: the directory in which the dataset is stored.
+  public init(
+    batchSize: Int, device: Device, entropy: Entropy, flattening: Bool = false, 
+    normalizing: Bool = false, 
+    localStorageDirectory: URL = DatasetUtilities.defaultDirectory
+      .appendingPathComponent("FashionMNIST", isDirectory: true)
+  ) {
+    training = TrainingEpochs(
+      samples: fetchMNISTDataset(
+        localStorageDirectory: localStorageDirectory,
+        remoteBaseDirectory: "http://fashion-mnist.s3-website.eu-central-1.amazonaws.com/",
+        imagesFilename: "train-images-idx3-ubyte",
+        labelsFilename: "train-labels-idx1-ubyte"),
+      batchSize: batchSize, entropy: entropy
+    ).lazy.map { (batches: Batches) -> LazyMapSequence<Batches, LabeledImage> in
+      return batches.lazy.map{ makeMNISTBatch(
+        samples: $0, flattening: flattening, normalizing: normalizing, device:device
+      )}
     }
+
+    validation = fetchMNISTDataset(
+      localStorageDirectory: localStorageDirectory,
+      remoteBaseDirectory: "http://fashion-mnist.s3-website.eu-central-1.amazonaws.com/",
+      imagesFilename: "t10k-images-idx3-ubyte",
+      labelsFilename: "t10k-labels-idx1-ubyte"
+    ).inBatches(of: batchSize).lazy.map {
+      makeMNISTBatch(samples: $0, flattening: flattening, normalizing: normalizing, 
+                     device:device)
+    }
+  }
 }
+
+extension FashionMNIST: ImageClassificationData where Entropy == SystemRandomNumberGenerator {
+  /// Creates an instance with `batchSize`.
+  public init(batchSize: Int) {
+    self.init(batchSize: batchSize, entropy: SystemRandomNumberGenerator())
+  }
+}