The network model to train our time series looks as follows:

sess = tf.Session() 
num_predictors = len(train_features.columns) 
num_classes = len(train_labels.columns) 
feature_data = tf.placeholder("float", [None, num_predictors]) 
actual_classes = tf.placeholder("float", [None, 2]) 
weights1 = tf.Variable(tf.truncated_normal([len(codes) * 3, 50], stddev=0.0001)) 
biases1 = tf.Variable(tf.ones([50])) 
weights2 = tf.Variable(tf.truncated_normal([50, 25], stddev=0.0001)) 
biases2 = tf.Variable(tf.ones([25])) 
weights3 = tf.Variable(tf.truncated_normal([25, 2], stddev=0.0001)) 
biases3 = tf.Variable(tf.ones([2])) 
hidden_layer_1 = tf.nn.relu(tf.matmul(feature_data, weights1) + biases1) 
hidden_layer_2 = tf.nn.relu(tf.matmul(hidden_layer_1, weights2) + biases2) 
model = tf.nn.softmax(tf.matmul(hidden_layer_2, weights3) + biases3) 
cost = -tf.reduce_sum(actual_classes * tf.log(model)) 
train_op1 = tf.train.AdamOptimizer(learning_rate=0.0001).minimize(cost) 
init = tf.initialize_all_variables() 
sess.run(init) 
correct_prediction = tf.equal(tf.argmax(model, 1), tf.argmax(actual_classes, 1)) 
accuracy = tf.reduce_mean(tf.cast(correct_prediction, "float")) 

This is just a simple network with two hidden layers.