Class: CooCoo::Trainer::Stochastic

Inherits:
Base show all
Defined in:
lib/coo-coo/trainer/stochastic.rb

Overview

Implements straight up stochastic gradient descent. No alterations get made to any hyperparameters while learning happens after every example.

Constant Summary

Constants inherited from Base

Base::DEFAULT_OPTIONS

Instance Method Summary collapse

Methods inherited from Base

#name, #options

Instance Method Details

#learn(network, input, expecting, rate, cost_function = CostFunctions::MeanSquare, hidden_state) ⇒ Object 



36
37
38
39
40
41
42
43
44
 
# File 'lib/coo-coo/trainer/stochastic.rb', line 36

def learn(network, input, expecting, rate, cost_function = CostFunctions::MeanSquare, hidden_state)
  output, hidden_state = network.forward(input, hidden_state)
  target = network.prep_output_target(expecting)
  final_output = network.final_output(output)
  errors = cost_function.derivative(target, final_output)
  deltas, hidden_state = network.backprop(input, output, errors, hidden_state)
  network.update_weights!(input, output, deltas * rate)
  return cost_function.call(target, final_output), hidden_state
end

#train(options, &block) ⇒ Object 



12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
 
# File 'lib/coo-coo/trainer/stochastic.rb', line 12

def train(options, &block)
  options = options.to_h
  network = options.fetch(:network)
  training_data = options.fetch(:data)
  learning_rate = options.fetch(:learning_rate, 0.3)
  batch_size = options.fetch(:batch_size, 1024)
  cost_function = options.fetch(:cost_function, CostFunctions::MeanSquare)
  
  t = Time.now
  
  training_data.each_slice(batch_size).with_index do |batch, i|
    total_errs = batch.inject(nil) do |acc, (expecting, input)|
      errs, hidden_state = learn(network, input, expecting, learning_rate, cost_function, Hash.new)
      errs + (acc || 0)
    end

    if block
      block.call(BatchStats.new(self, i, batch.size, Time.now - t, total_errs))
    end
    
    t = Time.now
  end
end