Class: TensorStream::Placeholder
- Defined in:
- lib/tensor_stream/placeholder.rb
Overview
Class that defines a TensorStream placeholder
Instance Attribute Summary
Attributes inherited from Tensor
#breakpoint, #data_type, #given_name, #graph, #internal, #is_const, #name, #native_buffer, #rank, #shape, #source, #value
Instance Method Summary collapse
-
#initialize(data_type, rank, shape, options = {}) ⇒ Placeholder
constructor
A new instance of Placeholder.
Methods inherited from Tensor
#!=, #*, #**, #+, #-, #-@, #/, #<, #<=, #==, #>, #>=, #[], #auto_math, #breakpoint!, cast_dtype, #collect, detect_type, #dtype, #eval, #first, #internal?, reset_counters, #to_a, #to_f, #to_h, #to_i, #to_math, #to_s
Methods included from OpHelper
#cons, #dtype_eval, #fp_type?, #i_cons, #i_op, #op, #shape_eval, #val_to_dtype
Constructor Details
#initialize(data_type, rank, shape, options = {}) ⇒ Placeholder
Returns a new instance of Placeholder.
4 5 6 7 8 9 10 11 12 13 14 15 16 |
# File 'lib/tensor_stream/placeholder.rb', line 4 def initialize(data_type, rank, shape, = {}) @graph = [:graph] || TensorStream.get_default_graph @data_type = data_type @rank = rank @shape = TensorShape.new(shape, rank) @value = nil @is_const = false @source = format_source(caller_locations) @name = [:name] || build_name @graph.add_node(self) end |