ruby-dnn 0.15.0 → 0.15.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 6be46c1c89fbb1f1f2091c95f15ccab21473d52571ee47953a4c07446ea8544c
-  data.tar.gz: e01eb929c13b2a33a350ee42b8bb8584446722ac3ba3dc7e163f7f3429ec068a
+  metadata.gz: 8fb316d982b142cfc31201f65e89b99579bdd1669d226c42dbc562c7af3b2c2a
+  data.tar.gz: 1839e28107f410d0008a21d9de3670fd5880e1a526ec24eddc34562e18390958
 SHA512:
-  metadata.gz: cc5bd6608d17a90cb97f6f5530877e5441a2dc73fc18e7b821c33d606469d6a15c22d519570b8c8821282915de84b677b3527221b1eb076a145a85997a58d2d9
-  data.tar.gz: 721c578f41a7648dfc5c462a1b9a81f6b9604b50540e8fa4a58bb801adc8059ee7cff9f27d3c1b1a710ce0857e426ca2087f4902bf46d9c621870e3ae7cb664a
+  metadata.gz: fd07fbd6409079cca07d14d08051e87b0e4b43772128f46f8cfe62857d2a1053bcb5dc13964f992bcb637ae64e7430613541bb19699d1fe90b142f9f950acfb8
+  data.tar.gz: 575171f1247cf0aa84fc71aee2b7318db10b51a0772e8dc1647e8189e1af70c39bd275edca1e0de8c9b45793051fe0fe8d75bce7bb6ec2e32f11359710592aa3

data/README.md

@@ -39,9 +39,9 @@ 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])
 ```
 
 When create a model with 'define by run' style:  
@@ -70,7 +70,7 @@ 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])
 ```
 
 Please refer to examples for basic usage.  
@@ -80,7 +80,7 @@ If you want to know more detailed information, please refer to the source code.
 || Implemented classes |
 |:-----------|------------:|
 | Connections | Dense, Conv2D, Conv2DTranspose, Embedding, SimpleRNN, LSTM, GRU |
-| Activations | Sigmoid, Tanh, Softsign, Softplus, Swish, ReLU, LeakyReLU, ELU |
+| Activations | Sigmoid, Tanh, Softsign, Softplus, Swish, ReLU, LeakyReLU, ELU, Mish |
 | Basic | Flatten, Reshape, Dropout, BatchNormalization |
 | Pooling | MaxPool2D, AvgPool2D, GlobalAvgPool2D, UnPool2D |
 | Optimizers | SGD, Nesterov, AdaGrad, RMSProp, AdaDelta, RMSPropGraves, Adam, AdaBound |

data/lib/dnn/core/layers/activations.rb

@@ -134,5 +134,18 @@ module DNN
       end
     end
 
+    class Mish < Layer
+      def forward(x)
+        @x = x
+        x * Xumo::NMath.tanh(Softplus.new.forward(x))
+      end
+
+      def backward(dy)
+        omega = 4 * (@x + 1) + 4 * Xumo::NMath.exp(2 * @x) + Xumo::NMath.exp(3 * @x) + Xumo::NMath.exp(@x) * (4 * @x + 6)
+        delta = 2 * Xumo::NMath.exp(@x) + Xumo::NMath.exp(2 * @x) + 2
+        dy * (Xumo::NMath.exp(@x) * omega) / delta**2
+      end
+    end
+
   end
 end

data/lib/dnn/core/models.rb

@@ -158,7 +158,7 @@ module DNN
 
       # Start training by iterator.
       # Setup the model before use this method.
-      # @param [Iterator] train_iterator Iterator used for training.
+      # @param [DNN::Iterator] train_iterator Iterator used for training.
       # @param [Integer] epochs Number of training.
       # @param [Integer] batch_size Batch size used for one training.
       # @param [Integer] initial_epoch Initial epoch.
@@ -260,10 +260,13 @@ module DNN
       # @return [Array] Returns the test data accuracy and mean loss in the form [accuracy, mean_loss].
       def evaluate(x, y, batch_size: 100)
         check_xy_type(x, y)
-        evaluate_by_iterator(Iterator.new(x, y, random: false))
+        evaluate_by_iterator(Iterator.new(x, y, random: false), batch_size: batch_size)
       end
 
-      # Evaluate model by iterator
+      # Evaluate model by iterator.
+      # @param [DNN::Iterator] test_iterator Iterator used for testing.
+      # @param [Integer] batch_size Batch size used for one test.
+      # @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
@@ -298,6 +301,7 @@ module DNN
       # Implement the process to accuracy this model.
       # @param [Numo::SFloat] x Input test data.
       # @param [Numo::SFloat] y Output test data.
+      # @return [Integer] Returns the test data accuracy.
       private def accuracy(x, y)
         if x.shape[1..-1] == [1]
           correct = 0

data/lib/dnn/core/utils.rb

@@ -30,5 +30,10 @@ module DNN
     def self.softmax(x)
       Losses::SoftmaxCrossEntropy.softmax(x)
     end
+
+    # Perform numerical differentiation.
+    def self.numerical_grad(x, func)
+      (func.(x + 1e-7) - func.(x)) / 1e-7
+    end
   end
 end

data/lib/dnn/version.rb

@@ -1,3 +1,3 @@
 module DNN
-  VERSION = "0.15.0"
+  VERSION = "0.15.1"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: ruby-dnn
 version: !ruby/object:Gem::Version
-  version: 0.15.0
+  version: 0.15.1
 platform: ruby
 authors:
 - unagiootoro
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2019-11-16 00:00:00.000000000 Z
+date: 2019-11-23 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: numo-narray