NMZivkovic

import numpy as np 
from matplotlib import pyplot as plt
from keras.utils.np_utils import to_categorical 
from keras.models import Sequential 
from keras.layers.core import Dense, Dropout, Flatten 
from keras.layers import Conv2D, MaxPooling2D

NMZivkovic

from keras.datasets import mnist

(X_train, y_train), (X_test, y_test) = mnist.load_data()

rows, cols = X_train[0].shape[0], X_train[0].shape[1] 
X_train = X_train.reshape(X_train.shape[0], rows, cols, 1) 
X_test = X_test.reshape(X_test.shape[0], rows, cols, 1) 

X_train = X_train.astype('float32')/255
X_test = X_test.astype('float32')/255

NMZivkovic

model = Sequential() 

model.add(Conv2D(32, kernel_size=(3, 3), activation='rectifier', input_shape=(rows, cols, 1)))
model.add(Conv2D(64, kernel_size=(3, 3), activation='rectifier')) 
model.add(Conv2D(128, kernel_size=(3, 3), activation='rectifier')) 

model.add(Dropout(0.5)) 

model.add(MaxPooling2D(pool_size = (2, 2)))

NMZivkovic

from keras.layers.advanced_activations import LeakyReLU

model.add(Conv2D(32, kernel_size=(3, 3), activation='linear',input_shape=(rows, cols, 1)))
model.add(LeakyReLU(alpha=0.1))

NMZivkovic

model.fit(X_train, y_train, batch_size=128, epochs=20, verbose=1, validation_split=0.2) 

score = model.evaluate(X_test, y_test, verbose=0) 

print('Accuracy:', score[1])

NMZivkovic

predictions = model.predict(X_test)

NMZivkovic

plt.figure(figsize=(15, 15)) 
for i in range(10):    
    ax = plt.subplot(2, 10, i + 1)    
    plt.imshow(X_test[i, :, :, 0], cmap='gray')    
    plt.title("Digit: {}\nPredicted:    {}".format(np.argmax(y_test[i]), np.argmax(predictions[i])))    
    plt.axis('off') 
plt.show()

NMZivkovic

import numpy as np

class RNN:
  
  def step(self, x):
    # Update the state
    self.h = np.tanh(np.dot(self.W_hh, self.h) + np.dot(self.W_xh, x))
    # Calculate the output
    y = np.dot(self.W_hy, self.h)
    return y

NMZivkovic

public class RNN
{
    private Matrix<double> _state;
    private Matrix<double> _inputWeights;
    private Matrix<double> _recurrentWeights;
    private Matrix<double> _outputWeights;

    public RNN(Matrix<double> initialInputWeights, Matrix<double> initialReccurentWeights, Matrix<double> initialOutputWeights)
    {
        _inputWeights = initialInputWeights;

NMZivkovic

import numpy as np
import collections

class DataHandler:
    def read_data(self, fname):
        with open(fname) as f:
            content = f.readlines()
        content = [x.strip() for x in content]
        content = [content[i].split() for i in range(len(content))]
        content = np.array(content)