Class: TensorStream::Variable
- Inherits:
-
Tensor
- Object
- Tensor
- TensorStream::Variable
show all
- Defined in:
- lib/tensor_stream/variable.rb
Overview
Class that defines a TensorStream variable
Instance Attribute Summary collapse
Attributes inherited from Tensor
#breakpoint, #consumers, #data_type, #device, #given_name, #graph, #internal, #is_const, #name, #native_buffer, #outputs, #rank, #shape, #source, #value
Class Method Summary
collapse
Instance Method Summary
collapse
Methods inherited from Tensor
#!=, #*, #**, #+, #-, #-@, #/, #<, #<=, #==, #>, #>=, #[], #and, #auto_math, #breakpoint!, cast_dtype, #collect, detect_type, #dot, #dtype, #eval, #first, #internal?, #matmul, #op, #print!, reset_counters, #to_a, #to_f, #to_h, #to_i, #to_s
Methods included from OpHelper
#_op, #cons, #dtype_eval, #format_source, #fp_type?, #i_cons, #i_op, #shape_eval, #val_to_dtype
Constructor Details
#initialize(data_type, rank, shape, options = {}) ⇒ Variable
Returns a new instance of Variable.
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
|
# File 'lib/tensor_stream/variable.rb', line 5
def initialize(data_type, rank, shape, options = {})
setup_initial_state(options)
@options = {
}
@data_type = data_type
@rank = rank
@value = nil
@is_const = false
@name = [TensorStream.get_variable_scope, options[:name] || build_name].compact.reject(&:empty?).join('/')
@initalizer_tensor = options[:initializer] ? options[:initializer] : _variable_scope.initializer || TensorStream.glorot_uniform_initializer
if shape.nil? && @initalizer_tensor && @initalizer_tensor.shape
shape = @initalizer_tensor.shape.shape
end
@shape = TensorShape.new(shape, rank)
@trainable = options.fetch(:trainable, true)
@graph.add_variable(self, options)
end
|
Instance Attribute Details
#buffer ⇒ Object
Returns the value of attribute buffer.
4
5
6
|
# File 'lib/tensor_stream/variable.rb', line 4
def buffer
@buffer
end
|
#options ⇒ Object
Returns the value of attribute options.
4
5
6
|
# File 'lib/tensor_stream/variable.rb', line 4
def options
@options
end
|
#trainable ⇒ Object
Returns the value of attribute trainable.
4
5
6
|
# File 'lib/tensor_stream/variable.rb', line 4
def trainable
@trainable
end
|
Class Method Details
.global_variables_initializer ⇒ Object
66
67
68
|
# File 'lib/tensor_stream/variable.rb', line 66
def self.global_variables_initializer
variables_initializer(TensorStream::GraphKeys::GLOBAL_VARIABLES)
end
|
.variables_initializer(collection) ⇒ Object
62
63
64
|
# File 'lib/tensor_stream/variable.rb', line 62
def self.variables_initializer(collection)
TensorStream.group(TensorStream.get_default_graph.get_collection(collection).map(&:initializer))
end
|
Instance Method Details
#assign(value, name: nil) ⇒ Object
35
36
37
38
|
# File 'lib/tensor_stream/variable.rb', line 35
def assign(value, name: nil)
_a, value = TensorStream.check_data_types(self, value)
Operation.new(:assign, self, value, name: name)
end
|
#assign_add(value) ⇒ Object
48
49
50
51
|
# File 'lib/tensor_stream/variable.rb', line 48
def assign_add(value)
_a, value = TensorStream.check_data_types(self, value)
Operation.new(:assign_add, self, value, data_type: data_type)
end
|
#assign_sub(value) ⇒ Object
57
58
59
60
|
# File 'lib/tensor_stream/variable.rb', line 57
def assign_sub(value)
_a, value = TensorStream.check_data_types(self, value)
Operation.new(:assign_sub, self, value)
end
|
#initializer ⇒ Object
28
29
30
31
32
33
|
# File 'lib/tensor_stream/variable.rb', line 28
def initializer
init_op = @initalizer_tensor.op
init_op.shape = @shape || init_op.shape
init_op.data_type = @data_type || init_op.data_type
assign(init_op)
end
|
#read_value ⇒ Object
40
41
42
43
44
45
46
|
# File 'lib/tensor_stream/variable.rb', line 40
def read_value
if buffer
@value = buffer.to_ruby
end
@value
end
|
#to_math(_tensor, _name_only = false, _max_depth = 99, _unused = 0) ⇒ Object
53
54
55
|
# File 'lib/tensor_stream/variable.rb', line 53
def to_math(_tensor, _name_only = false, _max_depth = 99, _unused = 0)
@name
end
|
#trainable? ⇒ Boolean
24
25
26
|
# File 'lib/tensor_stream/variable.rb', line 24
def trainable?
@trainable
end
|