In [1]:

    

import os
import sys

module_path = os.path.abspath(os.path.join('/home/felipe/neural-networks-and-deep-learning/src/'))
if module_path not in sys.path:
    sys.path.append(module_path)


    

In [2]:

    

import mnist_loader


    

In [4]:

    

training_data, validation_data, test_data = mnist_loader.load_data_wrapper()


    

In [5]:

    

import network2


    

In [6]:

    

net = network2.Network([784,30,10])


    

In [ ]:

    

net.SGD(training_data,
        30,10,0.1,lmbda=5.0,
        evaluation_data=validation_data,
        monitor_evaluation_accuracy=True)


    

In [8]:

    

net = network2.Network([784,30,30,10])
net.SGD(training_data,
        30,10,0.1,lmbda=5.0,
        evaluation_data=validation_data,
        monitor_evaluation_accuracy=True)


    

    




Epoch 0 training complete
Accuracy on evaluation data: 9182 / 10000

Epoch 1 training complete
Accuracy on evaluation data: 9411 / 10000

Epoch 2 training complete
Accuracy on evaluation data: 9515 / 10000

Epoch 3 training complete
Accuracy on evaluation data: 9530 / 10000

Epoch 4 training complete
Accuracy on evaluation data: 9573 / 10000

Epoch 5 training complete
Accuracy on evaluation data: 9574 / 10000

Epoch 6 training complete
Accuracy on evaluation data: 9616 / 10000

Epoch 7 training complete
Accuracy on evaluation data: 9637 / 10000

Epoch 8 training complete
Accuracy on evaluation data: 9622 / 10000

Epoch 9 training complete
Accuracy on evaluation data: 9631 / 10000

Epoch 10 training complete
Accuracy on evaluation data: 9656 / 10000

Epoch 11 training complete
Accuracy on evaluation data: 9662 / 10000

Epoch 12 training complete
Accuracy on evaluation data: 9653 / 10000

Epoch 13 training complete
Accuracy on evaluation data: 9640 / 10000

Epoch 14 training complete
Accuracy on evaluation data: 9670 / 10000

Epoch 15 training complete
Accuracy on evaluation data: 9648 / 10000

Epoch 16 training complete
Accuracy on evaluation data: 9673 / 10000

Epoch 17 training complete
Accuracy on evaluation data: 9673 / 10000

Epoch 18 training complete
Accuracy on evaluation data: 9677 / 10000

Epoch 19 training complete
Accuracy on evaluation data: 9656 / 10000

Epoch 20 training complete
Accuracy on evaluation data: 9669 / 10000

Epoch 21 training complete
Accuracy on evaluation data: 9693 / 10000

Epoch 22 training complete
Accuracy on evaluation data: 9694 / 10000

Epoch 23 training complete
Accuracy on evaluation data: 9694 / 10000

Epoch 24 training complete
Accuracy on evaluation data: 9697 / 10000

Epoch 25 training complete
Accuracy on evaluation data: 9665 / 10000

Epoch 26 training complete
Accuracy on evaluation data: 9669 / 10000

Epoch 27 training complete
Accuracy on evaluation data: 9698 / 10000

Epoch 28 training complete
Accuracy on evaluation data: 9692 / 10000

Epoch 29 training complete
Accuracy on evaluation data: 9683 / 10000







    
Out[8]:





([],
 [9182,
  9411,
  9515,
  9530,
  9573,
  9574,
  9616,
  9637,
  9622,
  9631,
  9656,
  9662,
  9653,
  9640,
  9670,
  9648,
  9673,
  9673,
  9677,
  9656,
  9669,
  9693,
  9694,
  9694,
  9697,
  9665,
  9669,
  9698,
  9692,
  9683],
 [],
 [])

In [ ]: