# counts number of 1's in a sequence of 10 steps, but outputs a runnint total at each step.

seq1b.py

from keras.models import Sequential
from keras.layers import LSTM, TimeDistributed, Dense, Activation

import numpy as np
from random import shuffle

# counts number of 1's in a sequence of 10 steps, but outputs a runnint total at each step.

NUM_EXAMPLES = 1000

train_input = ['{0:010b}'.format(i) for i in range(2**10)]
shuffle(train_input)
train_input = [map(int,i) for i in train_input]
ti  = []
for i in train_input:
    temp_list = []
    for j in i:
            temp_list.append([j])
    ti.append(np.array(temp_list))
train_input = ti

train_output = []
for i in train_input:
    l1 = []
    count = 0
    for j in i:
        if j[0] == 1:
            count+=1

        l2 = ([0]*11)
        l2[count]=1
        l1.append(l2)
    train_output.append(l1)

test_input = train_input[NUM_EXAMPLES:]
test_output = train_output[NUM_EXAMPLES:]
train_input = train_input[:NUM_EXAMPLES]
train_output = train_output[:NUM_EXAMPLES]

print "test and training data loaded"

model = Sequential()
model.add(LSTM(14,  return_sequences=True, input_shape=(10,1)))
model.add(TimeDistributed(Dense(11, activation='softmax')))


model.compile(loss='categorical_crossentropy', optimizer='sgd', metrics=['accuracy'])

model.fit(np.asarray(train_input), np.asarray(train_output), epochs=100, batch_size=128)

#loss_and_metrics = model.evaluate(test_input, test_output)


print model.predict(np.asarray([[[1],[0],[0],[1],[1],[0],[1],[1],[1],[0]]]))