ruby-dnn 0.15.1 → 0.15.2

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 8fb316d982b142cfc31201f65e89b99579bdd1669d226c42dbc562c7af3b2c2a
-  data.tar.gz: 1839e28107f410d0008a21d9de3670fd5880e1a526ec24eddc34562e18390958
+  metadata.gz: ebd331d481ec073102bb4b8c79a6ef280741deaf147a7908ef83f35603f784b9
+  data.tar.gz: c65399f9f29b65be2ffdc179047f58ae17d6b598328163fc45bb2da5f5f64751
 SHA512:
-  metadata.gz: fd07fbd6409079cca07d14d08051e87b0e4b43772128f46f8cfe62857d2a1053bcb5dc13964f992bcb637ae64e7430613541bb19699d1fe90b142f9f950acfb8
-  data.tar.gz: 575171f1247cf0aa84fc71aee2b7318db10b51a0772e8dc1647e8189e1af70c39bd275edca1e0de8c9b45793051fe0fe8d75bce7bb6ec2e32f11359710592aa3
+  metadata.gz: fb50c04b888f520d441de0643a6c9870be6df0c7e6d5c54f7c2b8e441e339723945f427cef67c51dc4d250bf0f02c20150e0ca47e1c13974fbe19d272f2301fb
+  data.tar.gz: 696f4fcbe01aa29f31ea08d32c7e399789ef841e22bfd77029701f132c10687ea81f25fe56e3044de95e4e8cd4c86b947fc32b257896f4dbcc3c125710d52ab8

data/examples/api-examples/early_stopping_example.rb

@@ -8,19 +8,18 @@ include DNN::Layers
 include DNN::Optimizers
 include DNN::Losses
 include DNN::Callbacks
-MNIST = DNN::MNIST
 
 EPOCHS = 3
 BATCH_SIZE = 128
 
-x_train, y_train = MNIST.load_train
-x_test, y_test = MNIST.load_test
+x_train, y_train = DNN::MNIST.load_train
+x_test, y_test = DNN::MNIST.load_test
 
-x_train = Numo::SFloat.cast(x_train).reshape(x_train.shape[0], 784)
-x_test = Numo::SFloat.cast(x_test).reshape(x_test.shape[0], 784)
+x_train = x_train.reshape(x_train.shape[0], 784)
+x_test = x_test.reshape(x_test.shape[0], 784)
 
-x_train /= 255
-x_test /= 255
+x_train = Numo::SFloat.cast(x_train) / 255
+x_test = Numo::SFloat.cast(x_test) / 255
 
 y_train = DNN::Utils.to_categorical(y_train, 10, Numo::SFloat)
 y_test = DNN::Utils.to_categorical(y_test, 10, Numo::SFloat)

data/examples/api-examples/initializer_example.rb

@@ -8,19 +8,18 @@ include DNN::Layers
 include DNN::Initializers
 include DNN::Optimizers
 include DNN::Losses
-MNIST = DNN::MNIST
 
 EPOCHS = 3
 BATCH_SIZE = 128
 
-x_train, y_train = MNIST.load_train
-x_test, y_test = MNIST.load_test
+x_train, y_train = DNN::MNIST.load_train
+x_test, y_test = DNN::MNIST.load_test
 
-x_train = Numo::SFloat.cast(x_train).reshape(x_train.shape[0], 784)
-x_test = Numo::SFloat.cast(x_test).reshape(x_test.shape[0], 784)
+x_train = x_train.reshape(x_train.shape[0], 784)
+x_test = x_test.reshape(x_test.shape[0], 784)
 
-x_train /= 255
-x_test /= 255
+x_train = Numo::SFloat.cast(x_train) / 255
+x_test = Numo::SFloat.cast(x_test) / 255
 
 y_train = DNN::Utils.to_categorical(y_train, 10, Numo::SFloat)
 y_test = DNN::Utils.to_categorical(y_test, 10, Numo::SFloat)

data/examples/api-examples/regularizer_example.rb

@@ -8,20 +8,19 @@ include DNN::Layers
 include DNN::Regularizers
 include DNN::Optimizers
 include DNN::Losses
-MNIST = DNN::MNIST
 
 EPOCHS = 3
 BATCH_SIZE = 128
 L2_LAMBDA = 0.01
 
-x_train, y_train = MNIST.load_train
-x_test, y_test = MNIST.load_test
+x_train, y_train = DNN::MNIST.load_train
+x_test, y_test = DNN::MNIST.load_test
 
-x_train = Numo::SFloat.cast(x_train).reshape(x_train.shape[0], 784)
-x_test = Numo::SFloat.cast(x_test).reshape(x_test.shape[0], 784)
+x_train = x_train.reshape(x_train.shape[0], 784)
+x_test = x_test.reshape(x_test.shape[0], 784)
 
-x_train /= 255
-x_test /= 255
+x_train = Numo::SFloat.cast(x_train) / 255
+x_test = Numo::SFloat.cast(x_test) / 255
 
 y_train = DNN::Utils.to_categorical(y_train, 10, Numo::SFloat)
 y_test = DNN::Utils.to_categorical(y_test, 10, Numo::SFloat)

data/examples/api-examples/save_example.rb

@@ -9,7 +9,6 @@ include DNN::Optimizers
 include DNN::Losses
 include DNN::Savers
 include DNN::Loaders
-MNIST = DNN::MNIST
 
 USE_MARSHAL = 0
 USE_JSON = 1
@@ -24,14 +23,14 @@ SAVE_STYLE = USE_MARSHAL
 # This setting is enabled when SAVE_STYLE is USE_MARSHAL.
 INCLUDE_MODEL = true
 
-x_train, y_train = MNIST.load_train
-x_test, y_test = MNIST.load_test
+x_train, y_train = DNN::MNIST.load_train
+x_test, y_test = DNN::MNIST.load_test
 
-x_train = Numo::SFloat.cast(x_train).reshape(x_train.shape[0], 784)
-x_test = Numo::SFloat.cast(x_test).reshape(x_test.shape[0], 784)
+x_train = x_train.reshape(x_train.shape[0], 784)
+x_test = x_test.reshape(x_test.shape[0], 784)
 
-x_train /= 255
-x_test /= 255
+x_train = Numo::SFloat.cast(x_train) / 255
+x_test = Numo::SFloat.cast(x_test) / 255
 
 y_train = DNN::Utils.to_categorical(y_train, 10, Numo::SFloat)
 y_test = DNN::Utils.to_categorical(y_test, 10, Numo::SFloat)

data/examples/cifar100_example.rb

@@ -7,19 +7,12 @@ include DNN::Models
 include DNN::Layers
 include DNN::Optimizers
 include DNN::Losses
-CIFAR100 = DNN::CIFAR100
 
-x_train, y_train = CIFAR100.load_train
-x_test, y_test = CIFAR100.load_test
+x_train, y_train = DNN::CIFAR100.load_train
+x_test, y_test = DNN::CIFAR100.load_test
 
-x_train = Numo::SFloat.cast(x_train)
-x_test = Numo::SFloat.cast(x_test)
-
-x_train /= 255
-x_test /= 255
-
-y_train = y_train[true, 1]
-y_test = y_test[true, 1]
+x_train = Numo::SFloat.cast(x_train) / 255
+x_test = Numo::SFloat.cast(x_test) / 255
 
 y_train = DNN::Utils.to_categorical(y_train, 100, Numo::SFloat)
 y_test = DNN::Utils.to_categorical(y_test, 100, Numo::SFloat)
@@ -28,43 +21,44 @@ model = Sequential.new
 
 model << InputLayer.new([32, 32, 3])
 
-model << Conv2D.new(16, 5, padding: true)
-model << BatchNormalization.new
+model << Conv2D.new(32, 3, padding: true)
+model << Dropout.new(0.25)
 model << ReLU.new
 
-model << Conv2D.new(16, 5, padding: true)
+model << Conv2D.new(32, 3, padding: true)
 model << BatchNormalization.new
 model << ReLU.new
-
 model << MaxPool2D.new(2)
 
-model << Conv2D.new(32, 5, padding: true)
-model << BatchNormalization.new
+model << Conv2D.new(64, 3, padding: true)
+model << Dropout.new(0.25)
 model << ReLU.new
 
-model << Conv2D.new(32, 5, padding: true)
+model << Conv2D.new(64, 3, padding: true)
 model << BatchNormalization.new
 model << ReLU.new
-
 model << MaxPool2D.new(2)
 
-model << Conv2D.new(64, 5, padding: true)
-model << BatchNormalization.new
+model << Conv2D.new(128, 3, padding: true)
+model << Dropout.new(0.25)
 model << ReLU.new
 
-model << Conv2D.new(64, 5, padding: true)
+model << Conv2D.new(128, 3, padding: true)
 model << BatchNormalization.new
 model << ReLU.new
 
 model << Flatten.new
 
-model << Dense.new(1024)
+model << Dense.new(512)
 model << BatchNormalization.new
 model << ReLU.new
-model << Dropout.new(0.5)
 
 model << Dense.new(100)
 
 model.setup(Adam.new, SoftmaxCrossEntropy.new)
 
-model.train(x_train, y_train, 10, batch_size: 100, test: [x_test, y_test])
+model.train(x_train, y_train, 10, batch_size: 128, test: [x_test, y_test])
+
+accuracy, loss = model.evaluate(x_test, y_test)
+puts "accuracy: #{accuracy}"
+puts "loss: #{loss}"

data/examples/cifar10_example.rb

@@ -7,16 +7,12 @@ include DNN::Models
 include DNN::Layers
 include DNN::Optimizers
 include DNN::Losses
-CIFAR10 = DNN::CIFAR10
 
-x_train, y_train = CIFAR10.load_train
-x_test, y_test = CIFAR10.load_test
+x_train, y_train = DNN::CIFAR10.load_train
+x_test, y_test = DNN::CIFAR10.load_test
 
-x_train = Numo::SFloat.cast(x_train)
-x_test = Numo::SFloat.cast(x_test)
-
-x_train /= 255
-x_test /= 255
+x_train = Numo::SFloat.cast(x_train) / 255
+x_test = Numo::SFloat.cast(x_test) / 255
 
 y_train = DNN::Utils.to_categorical(y_train, 10, Numo::SFloat)
 y_test = DNN::Utils.to_categorical(y_test, 10, Numo::SFloat)
@@ -25,31 +21,29 @@ model = Sequential.new
 
 model << InputLayer.new([32, 32, 3])
 
-model << Conv2D.new(16, 5, padding: true)
-model << BatchNormalization.new
+model << Conv2D.new(32, 3, padding: true)
+model << Dropout.new(0.25)
 model << ReLU.new
 
-model << Conv2D.new(16, 5, padding: true)
+model << Conv2D.new(32, 3, padding: true)
 model << BatchNormalization.new
 model << ReLU.new
-
 model << MaxPool2D.new(2)
 
-model << Conv2D.new(32, 5, padding: true)
-model << BatchNormalization.new
+model << Conv2D.new(64, 3, padding: true)
+model << Dropout.new(0.25)
 model << ReLU.new
 
-model << Conv2D.new(32, 5, padding: true)
+model << Conv2D.new(64, 3, padding: true)
 model << BatchNormalization.new
 model << ReLU.new
-
 model << MaxPool2D.new(2)
 
-model << Conv2D.new(64, 5, padding: true)
-model << BatchNormalization.new
+model << Conv2D.new(128, 3, padding: true)
+model << Dropout.new(0.25)
 model << ReLU.new
 
-model << Conv2D.new(64, 5, padding: true)
+model << Conv2D.new(128, 3, padding: true)
 model << BatchNormalization.new
 model << ReLU.new
 
@@ -58,10 +52,13 @@ model << Flatten.new
 model << Dense.new(512)
 model << BatchNormalization.new
 model << ReLU.new
-model << Dropout.new(0.5)
 
 model << Dense.new(10)
 
 model.setup(Adam.new, SoftmaxCrossEntropy.new)
 
-model.train(x_train, y_train, 10, batch_size: 100, test: [x_test, y_test])
+model.train(x_train, y_train, 10, batch_size: 128, test: [x_test, y_test])
+
+accuracy, loss = model.evaluate(x_test, y_test)
+puts "accuracy: #{accuracy}"
+puts "loss: #{loss}"

data/examples/dcgan/train.rb

@@ -6,7 +6,6 @@ require_relative "dcgan"
 include DNN::Optimizers
 include DNN::Losses
 include DNN::Callbacks
-MNIST = DNN::MNIST
 
 Numo::SFloat.srand(rand(1 << 31))
 
@@ -21,7 +20,7 @@ dis.setup(Adam.new(alpha: 0.00001, beta1: 0.1), SigmoidCrossEntropy.new)
 dcgan.setup(Adam.new(alpha: 0.0002, beta1: 0.5), SigmoidCrossEntropy.new)
 dcgan.add_callback(CheckPoint.new("trained/dcgan_model"))
 
-x_train, * = MNIST.load_train
+x_train, * = DNN::MNIST.load_train
 x_train = Numo::SFloat.cast(x_train)
 x_train = x_train / 127.5 - 1
 

data/examples/iris_example.rb

@@ -7,9 +7,8 @@ include DNN::Models
 include DNN::Layers
 include DNN::Optimizers
 include DNN::Losses
-Iris = DNN::Iris
 
-x, y = Iris.load(true)
+x, y = DNN::Iris.load(true)
 x_train, y_train = x[0...100, true], y[0...100]
 x_test, y_test = x[100...150, true], y[100...150]
 
@@ -27,4 +26,8 @@ model << Dense.new(3)
 
 model.setup(Adam.new, SoftmaxCrossEntropy.new)
 
-model.train(x_train, y_train, 1000, batch_size: 10, test: [x_test, y_test])
+model.train(x_train, y_train, 500, batch_size: 32, test: [x_test, y_test])
+
+accuracy, loss = model.evaluate(x_test, y_test)
+puts "accuracy: #{accuracy}"
+puts "loss: #{loss}"

data/examples/mnist_conv2d_example.rb

@@ -7,16 +7,12 @@ include DNN::Models
 include DNN::Layers
 include DNN::Optimizers
 include DNN::Losses
-MNIST = DNN::MNIST
 
-x_train, y_train = MNIST.load_train
-x_test, y_test = MNIST.load_test
+x_train, y_train = DNN::MNIST.load_train
+x_test, y_test = DNN::MNIST.load_test
 
-x_train = Numo::SFloat.cast(x_train)
-x_test = Numo::SFloat.cast(x_test)
-
-x_train /= 255
-x_test /= 255
+x_train = Numo::SFloat.cast(x_train) / 255
+x_test = Numo::SFloat.cast(x_test) / 255
 
 y_train = DNN::Utils.to_categorical(y_train, 10, Numo::SFloat)
 y_test = DNN::Utils.to_categorical(y_test, 10, Numo::SFloat)
@@ -47,3 +43,7 @@ model << Dense.new(10)
 model.setup(Adam.new, SoftmaxCrossEntropy.new)
 
 model.train(x_train, y_train, 10, batch_size: 100, test: [x_test, y_test])
+
+accuracy, loss = model.evaluate(x_test, y_test)
+puts "accuracy: #{accuracy}"
+puts "loss: #{loss}"

data/examples/mnist_define_by_run.rb

@@ -1,22 +1,21 @@
 require "dnn"
 require "dnn/datasets/mnist"
 # If you use numo/linalg then please uncomment out.
-# require "numo/linalg/autoloader"
+require "numo/linalg/autoloader"
 
 include DNN::Models
 include DNN::Layers
 include DNN::Optimizers
 include DNN::Losses
-MNIST = DNN::MNIST
 
-x_train, y_train = MNIST.load_train
-x_test, y_test = MNIST.load_test
+x_train, y_train = DNN::MNIST.load_train
+x_test, y_test = DNN::MNIST.load_test
 
-x_train = Numo::SFloat.cast(x_train).reshape(x_train.shape[0], 784)
-x_test = Numo::SFloat.cast(x_test).reshape(x_test.shape[0], 784)
+x_train = x_train.reshape(x_train.shape[0], 784)
+x_test = x_test.reshape(x_test.shape[0], 784)
 
-x_train /= 255
-x_test /= 255
+x_train = Numo::SFloat.cast(x_train) / 255
+x_test = Numo::SFloat.cast(x_test) / 255
 
 y_train = DNN::Utils.to_categorical(y_train, 10, Numo::SFloat)
 y_test = DNN::Utils.to_categorical(y_test, 10, Numo::SFloat)
@@ -27,17 +26,13 @@ class MLP < Model
     @l1 = Dense.new(256)
     @l2 = Dense.new(256)
     @l3 = Dense.new(10)
-    @bn1 = BatchNormalization.new
-    @bn2 = BatchNormalization.new
   end
 
   def call(x)
     x = InputLayer.new(784).(x)
     x = @l1.(x)
-    x = @bn1.(x)
     x = ReLU.(x)
     x = @l2.(x)
-    x = @bn2.(x)
     x = ReLU.(x)
     x = @l3.(x)
     x
@@ -45,7 +40,10 @@ class MLP < Model
 end
 
 model = MLP.new
-
 model.setup(Adam.new, SoftmaxCrossEntropy.new)
 
-model.train(x_train, y_train, 10, batch_size: 100, test: [x_test, y_test])
+model.train(x_train, y_train, 10, batch_size: 128, test: [x_test, y_test])
+
+accuracy, loss = model.evaluate(x_test, y_test)
+puts "accuracy: #{accuracy}"
+puts "loss: #{loss}"

data/examples/mnist_example.rb

@@ -7,16 +7,15 @@ include DNN::Models
 include DNN::Layers
 include DNN::Optimizers
 include DNN::Losses
-MNIST = DNN::MNIST
 
-x_train, y_train = MNIST.load_train
-x_test, y_test = MNIST.load_test
+x_train, y_train = DNN::MNIST.load_train
+x_test, y_test = DNN::MNIST.load_test
 
-x_train = Numo::SFloat.cast(x_train).reshape(x_train.shape[0], 784)
-x_test = Numo::SFloat.cast(x_test).reshape(x_test.shape[0], 784)
+x_train = x_train.reshape(x_train.shape[0], 784)
+x_test = x_test.reshape(x_test.shape[0], 784)
 
-x_train /= 255
-x_test /= 255
+x_train = Numo::SFloat.cast(x_train) / 255
+x_test = Numo::SFloat.cast(x_test) / 255
 
 y_train = DNN::Utils.to_categorical(y_train, 10, Numo::SFloat)
 y_test = DNN::Utils.to_categorical(y_test, 10, Numo::SFloat)
@@ -33,6 +32,10 @@ model << ReLU.new
 
 model << Dense.new(10)
 
-model.setup(RMSProp.new, SoftmaxCrossEntropy.new)
+model.setup(Adam.new, SoftmaxCrossEntropy.new)
 
-model.train(x_train, y_train, 10, batch_size: 100, test: [x_test, y_test])
+model.train(x_train, y_train, 10, batch_size: 128, test: [x_test, y_test])
+
+accuracy, loss = model.evaluate(x_test, y_test)
+puts "accuracy: #{accuracy}"
+puts "loss: #{loss}"

data/examples/mnist_lstm_example.rb

@@ -7,16 +7,15 @@ include DNN::Models
 include DNN::Layers
 include DNN::Optimizers
 include DNN::Losses
-MNIST = DNN::MNIST
 
-x_train, y_train = MNIST.load_train
-x_test, y_test = MNIST.load_test
+x_train, y_train = DNN::MNIST.load_train
+x_test, y_test = DNN::MNIST.load_test
 
-x_train = Numo::SFloat.cast(x_train).reshape(x_train.shape[0], 28, 28)
-x_test = Numo::SFloat.cast(x_test).reshape(x_test.shape[0], 28, 28)
+x_train = x_train.reshape(x_train.shape[0], 28, 28)
+x_test = x_test.reshape(x_test.shape[0], 28, 28)
 
-x_train /= 255
-x_test /= 255
+x_train = Numo::SFloat.cast(x_train) / 255
+x_test = Numo::SFloat.cast(x_test) / 255
 
 y_train = DNN::Utils.to_categorical(y_train, 10, Numo::SFloat)
 y_test = DNN::Utils.to_categorical(y_test, 10, Numo::SFloat)
@@ -33,3 +32,7 @@ model << Dense.new(10)
 model.setup(Adam.new, SoftmaxCrossEntropy.new)
 
 model.train(x_train, y_train, 10, batch_size: 100, test: [x_test, y_test])
+
+accuracy, loss = model.evaluate(x_test, y_test)
+puts "accuracy: #{accuracy}"
+puts "loss: #{loss}"

data/lib/dnn/core/models.rb

@@ -269,7 +269,7 @@ module DNN
       # @return [Array] Returns the test data accuracy and mean loss in the form [accuracy, mean_loss].
       def evaluate_by_iterator(test_iterator, batch_size: 100)
         num_test_datas = test_iterator.num_datas
-        batch_size = batch_size >= num_test_datas[0] ? num_test_datas : batch_size
+        batch_size = batch_size >= num_test_datas ? num_test_datas : batch_size
         total_correct = 0
         sum_loss = 0
         max_steps = (num_test_datas.to_f / batch_size).ceil

data/lib/dnn/version.rb

@@ -1,3 +1,3 @@
 module DNN
-  VERSION = "0.15.1"
+  VERSION = "0.15.2"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: ruby-dnn
 version: !ruby/object:Gem::Version
-  version: 0.15.1
+  version: 0.15.2
 platform: ruby
 authors:
 - unagiootoro
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2019-11-23 00:00:00.000000000 Z
+date: 2019-12-01 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: numo-narray