tensor_stream 0.9.8 → 0.9.9

data/lib/tensor_stream/variable.rb

@@ -1,16 +1,21 @@
 module TensorStream
   # Class that defines a TensorStream variable
   class Variable < Tensor
-    attr_accessor :trainable, :options, :buffer
-    def initialize(data_type, rank, shape, variable_scope, options = {})
-      setup_initial_state(options)
+    attr_accessor :trainable, :options, :buffer, :op
+    attr_writer :value
 
-      @options = {
-      }
+    def initialize(data_type)
       @data_type = data_type
+      @options = {}
+      @is_const = false
+    end
+
+    def prepare(rank, shape, variable_scope, options = {})
+      setup_initial_state(options)
+
       @rank = rank
       @value = nil
-      @is_const = false
+
       scope_name = variable_scope ? variable_scope.name : nil
       variable_scope_initializer = variable_scope ? variable_scope.initializer : nil
       @name = [scope_name, options[:name] || build_name].compact.reject(&:empty?).join('/')
@@ -19,7 +24,6 @@ module TensorStream
 
       @shape = TensorShape.new(shape, rank)
       @trainable = options.fetch(:trainable, true)
-      @graph.add_variable(self, options)
     end
 
     def trainable?
@@ -55,10 +59,6 @@ module TensorStream
       _op(:assign_add, self, value, data_type: data_type, name: name)
     end
 
-    def op
-      @op ||= _op(:variable, self, data_type: data_type)
-    end
-
     def to_math(_tensor, _name_only = false, _max_depth = 99, _unused = 0)
       @name
     end
@@ -75,5 +75,11 @@ module TensorStream
     def self.global_variables_initializer
       variables_initializer(TensorStream::GraphKeys::GLOBAL_VARIABLES)
     end
+
+    protected
+
+    def build_name
+      "Variable#{graph.get_var_counter}:#{@rank}"
+    end
   end
 end

data/lib/tensor_stream/variable_scope.rb

@@ -1,5 +1,6 @@
 module TensorStream
   class VariableScope
+    include OpHelper
     attr_accessor :name, :reuse, :initializer
     attr_reader :used_names
 
@@ -12,7 +13,8 @@ module TensorStream
 
     def get_variable(name, dtype: nil, shape: nil, initializer: nil, trainable: true, collections: nil, validate_shape: false)
       raise TensorStream::ValueError, "validate_shape=true and initializer does not have a defined shape" if validate_shape && !shape.nil && initializer.is_a?(Tensor)
-      TensorStream::Variable.new(dtype || :float32, nil, shape, self, collections: collections, name: name, initializer: initializer, trainable: trainable)
+
+      i_var(dtype || :float32, nil, shape, self, collections: collections, name: name, initializer: initializer, trainable: trainable)
     end
 
     def register_name(name)

data/lib/tensor_stream/version.rb

@@ -1,5 +1,5 @@
 module TensorStream
-  VERSION = '0.9.8'.freeze
+  VERSION = '0.9.9'.freeze
 
   def self.version
     VERSION

data/samples/iris.rb

@@ -52,15 +52,14 @@ end
 
 def init_weights(shape)
   # Weight initialization
-  weights = TensorStream.random_normal(shape, stddev: 0.1)
-  TensorStream.variable(weights)
+  TensorStream.random_normal(shape, stddev: 0.1).var
 end
 
 def forwardprop(x, w_1, w_2)
   # Forward-propagation.
   # IMPORTANT: yhat is not softmax since TensorFlow's softmax_cross_entropy_with_logits() does that internally.
-  h  = TensorStream.nn.sigmoid(TensorStream.matmul(x, w_1))  # The \sigma function
-  TensorStream.matmul(h, w_2)  # The \varphi function
+  h  = TensorStream.nn.sigmoid(x.matmul w_1)  # The \sigma function
+  h.matmul w_2  # The \varphi function
 end
 
 x_size = x_train[0].size

data/samples/linear_regression.rb

@@ -16,19 +16,19 @@ train_Y = [1.7,2.76,2.09,3.19,1.694,1.573,3.366,2.596,2.53,1.221,
 
 n_samples = train_X.size
 
-X = tf.placeholder("float")
-Y = tf.placeholder("float")
+X = Float.placeholder
+Y = Float.placeholder
 
 # Set model weights
 
-W = tf.variable(rand, name: "weight")
-b = tf.variable(rand, name: "bias")
+W = rand.t.var name: "weight"
+b = rand.t.var name: "bias"
 
 # Construct a linear model
 pred = X * W + b
 
 # Mean squared error
-cost = ((pred - Y) ** 2).reduce(:+) / ( 2 * n_samples)
+cost = ((pred - Y) ** 2).reduce / ( 2 * n_samples)
 
 # Other possible Optimizers
 
@@ -42,6 +42,8 @@ optimizer = TensorStream::Train::GradientDescentOptimizer.new(learning_rate).min
 
 # Initialize the variables (i.e. assign their default value)
 init = tf.global_variables_initializer()
+# Add ops to save and restore all the variables.
+saver = tf::Train::Saver.new
 
 tf.session do |sess|
     start_time = Time.now
@@ -52,6 +54,8 @@ tf.session do |sess|
       end
 
       if (epoch+1) % display_step == 0
+        # Save the variables to disk.
+        save_path = saver.save(sess, "/tmp/lg_model")
         c = sess.run(cost, feed_dict: {X => train_X, Y => train_Y})
         puts("Epoch:", '%04d' % (epoch+1), "cost=",  c, \
             "W=", sess.run(W), "b=", sess.run(b))

data/samples/logistic_regression.rb

@@ -39,19 +39,19 @@ test_x = transformed_data[51..100].map { |x| x[0..3].map(&:to_f) }
 test_y = iris[51..100].map { |x| x[4] == 'Iris-setosa' ? 0.0 : 1.0 }
 
 
-A = tf.variable(tf.random_normal([4, 1]))
-b = tf.variable(tf.random_normal([1, 1]))
+A = tf.random_normal([4, 1]).var
+b = tf.random_normal([1, 1]).var
 
 init = tf.global_variables_initializer
 sess = tf.session
 sess.run(init)
 
-data = tf.placeholder(:float32, shape: [nil, 4])
-target = tf.placeholder(:float32, shape: [nil, 1])
+data = Float.placeholder shape: [nil, 4]
+target = Float.placeholder shape: [nil, 1]
 
-mod = data.dot(A) + b
+mod = data.matmul(A) + b
 
-loss = tf.reduce_mean(tf.nn.sigmoid_cross_entropy_with_logits(logits: mod, labels: target))
+loss = tf.nn.sigmoid_cross_entropy_with_logits(logits: mod, labels: target).reduce :mean
 
 learning_rate = 0.003
 batch_size = 30
@@ -59,11 +59,13 @@ iter_num = 1500
 
 optimizer = TensorStream::Train::GradientDescentOptimizer.new(learning_rate)
 goal = optimizer.minimize(loss)
-prediction = tf.round(tf.sigmoid(mod))
+prediction = tf.sigmoid(mod).round
+
 # Bool into float32 type
-correct = tf.cast(tf.equal(prediction, target), :float32)
+correct = (prediction == target).cast
+
 # Average
-accuracy = tf.reduce_mean(correct)
+accuracy = correct.reduce :mean
 
 loss_trace = []
 train_acc = []

data/samples/mnist_data.rb

@@ -20,22 +20,20 @@ puts "downloading minst data"
 mnist = Mnist.read_data_sets('/tmp/data', one_hot: true)
 puts "downloading finished"
 
-x = tf.placeholder(:float32, shape: [nil, 784])
-w = tf.variable(tf.zeros([784, 10]))
-b = tf.variable(tf.zeros([10]))
-
-
+x = Float.placeholder shape: [nil, 784]
+w = tf.zeros([784, 10]).var
+b = tf.zeros([10]).var
 
 # model
-y = tf.nn.softmax(tf.matmul(tf.reshape(x, [-1, 784]), w) + b)
+y = tf.nn.softmax(x.reshape([-1, 784]).matmul(w) + b)
 
-y_ = tf.placeholder(:float32, shape: [nil, 10])
+y_ = Float.placeholder shape: [nil, 10]
 
 # loss function
-cross_entropy = -tf.reduce_sum(y_ * tf.log(y))
+cross_entropy = -(y_ * y.log).reduce
 
-is_correct = tf.equal(tf.argmax(y, 1), tf.argmax(y_, 1))
-accuracy =  tf.reduce_mean(tf.cast(is_correct, :float32))
+is_correct = tf.argmax(y, 1) == tf.argmax(y_, 1)
+accuracy = is_correct.cast.reduce :mean
 
 optimizer = TensorStream::Train::AdamOptimizer.new
 train_step = optimizer.minimize(cross_entropy)

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: tensor_stream
 version: !ruby/object:Gem::Version
-  version: 0.9.8
+  version: 0.9.9
 platform: ruby
 authors:
 - Joseph Emmanuel Dayo
 autorequire: 
 bindir: exe
 cert_chain: []
-date: 2018-11-25 00:00:00.000000000 Z
+date: 2018-12-28 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -254,6 +254,7 @@ files:
 - data_expected.json
 - data_input.json
 - lib/tensor_stream.rb
+- lib/tensor_stream/constant.rb
 - lib/tensor_stream/control_flow.rb
 - lib/tensor_stream/debugging/debugging.rb
 - lib/tensor_stream/device.rb
@@ -275,14 +276,17 @@ files:
 - lib/tensor_stream/graph.rb
 - lib/tensor_stream/graph_builder.rb
 - lib/tensor_stream/graph_deserializers/protobuf.rb
+- lib/tensor_stream/graph_deserializers/yaml_loader.rb
 - lib/tensor_stream/graph_keys.rb
 - lib/tensor_stream/graph_serializers/graphml.rb
 - lib/tensor_stream/graph_serializers/packer.rb
 - lib/tensor_stream/graph_serializers/pbtext.rb
 - lib/tensor_stream/graph_serializers/serializer.rb
+- lib/tensor_stream/graph_serializers/yaml.rb
 - lib/tensor_stream/helpers/infer_shape.rb
 - lib/tensor_stream/helpers/op_helper.rb
 - lib/tensor_stream/helpers/string_helper.rb
+- lib/tensor_stream/helpers/tensor_mixins.rb
 - lib/tensor_stream/images.rb
 - lib/tensor_stream/initializer.rb
 - lib/tensor_stream/math_gradients.rb
@@ -312,6 +316,7 @@ files:
 - lib/tensor_stream/trainer.rb
 - lib/tensor_stream/types.rb
 - lib/tensor_stream/utils.rb
+- lib/tensor_stream/utils/freezer.rb
 - lib/tensor_stream/variable.rb
 - lib/tensor_stream/variable_scope.rb
 - lib/tensor_stream/version.rb
@@ -347,8 +352,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubyforge_project: 
-rubygems_version: 2.7.7
+rubygems_version: 3.0.1
 signing_key: 
 specification_version: 4
 summary: A Pure ruby tensorflow implementation