An implementation of a Neural Network used for ATR (Automatic Target Recognition)
In [5]:
# dependencies
import tensorflow as tf
import numpy as np
import pickle
In [6]:
# load data
pickle_file = 'final_dataset.pickle'
with open(pickle_file, 'rb') as f:
save = pickle.load(f)
train_dataset = save['train_dataset']
train_labels = save['train_labels']
valid_dataset = save['valid_dataset']
valid_labels = save['valid_labels']
test_dataset = save['test_dataset']
test_labels = save['test_labels']
del save # help gc free up memory
print('Training set', train_dataset.shape, train_labels.shape)
print('Validation set', valid_dataset.shape, valid_labels.shape)
print('Test set', test_dataset.shape, test_labels.shape)
In [7]:
# Now lets test if the file really matches or is corrupted
# train_labels[0] => 2
# so the first image is a BTR70 LETS test this out
print(train_labels[0])
with open('TRAIN_BTR70.pickle','rb') as f:
s = pickle.load(f)
btr_train = s
del s
for image in btr_train:
if (image - train_dataset[0]).any() == 0:
print('no problem')
break
print('done')
Reformat Data - Flatten arrays and make labels 1-hot encoded arrays
In [8]:
image_size = 128
num_labels = 3
def reformat(dataset, labels):
dataset = dataset.reshape((-1, image_size * image_size)).astype(np.float32)
# Map 0 to [1.0, 0.0, 0.0 ...], 1 to [0.0, 1.0, 0.0 ...]
labels = (np.arange(num_labels) == labels[:,None]).astype(np.float32)
return dataset, labels
train_dataset, train_labels = reformat(train_dataset, train_labels)
valid_dataset, valid_labels = reformat(valid_dataset, valid_labels)
test_dataset, test_labels = reformat(test_dataset, test_labels)
print('Training set', train_dataset.shape, train_labels.shape)
print('Validation set', valid_dataset.shape, valid_labels.shape)
print('Test set', test_dataset.shape, test_labels.shape)
In [105]:
# hyper parameters
num_steps = 551
batch_size = 30
num_labels = 3
h_nodes = 200
beta = 0.01
In [10]:
def accuracy(predictions, labels):
return (100.0 * np.sum(np.argmax(predictions, 1) == np.argmax(labels, 1))
/ predictions.shape[0])
In [109]:
train_subset = 30
graph = tf.Graph()
with graph.as_default():
# Input data.
# Load the training, validation and test data into constants that are
# attached to the graph.
# in tensorflow you create a bunch of nodes or operations - some are constant (do not require tensor input)
# and some are not constant example matrix multilication -the end node that you want as output is supposed
# to be passed as a parameter to the session variable
#placing inside constant means that you have do not perform any computation on these tensors
# everything is an operation the below one produces a matrix
tf_train_dataset = tf.constant(train_dataset[:train_subset, :])
tf_train_labels = tf.constant(train_labels[:train_subset])
tf_valid_dataset = tf.constant(valid_dataset)
tf_test_dataset = tf.constant(test_dataset)
# Variables.
# These are the parameters that we are going to be training. The weight
# matrix will be initialized using random values following a (truncated)
# normal distribution. The biases get initialized to zero.
weights = tf.Variable(
tf.truncated_normal([image_size * image_size, num_labels]))
biases = tf.Variable(tf.zeros([num_labels]))
# Training computation.
# We multiply the inputs with the weight matrix, and add biases. We compute
# the softmax and cross-entropy (it's one operation in TensorFlow, because
# it's very common, and it can be optimized). We take the average of this
# cross-entropy across all training examples: that's our loss.
# y = (W*x) + b
logits = tf.matmul(tf_train_dataset, weights) + biases
# S(y)-> will be reduced to one hot encoded values then cross entropy will be calculated
# the log function D(S,L) that is the loss
loss = tf.reduce_mean(
tf.nn.softmax_cross_entropy_with_logits(labels=tf_train_labels, logits=logits))
# lets add the l2 regularization layer
regularization = tf.nn.l2_loss(weights)
loss = tf.reduce_mean(loss + beta*regularization)
# Optimizer.
# We are going to find the minimum of this loss using gradient descent.
optimizer = tf.train.GradientDescentOptimizer(0.5).minimize(loss)
# Predictions for the training, validation, and test data.
# These are not part of training, but merely here so that we can report
# accuracy figures as we train.
train_prediction = tf.nn.softmax(logits)
valid_prediction = tf.nn.softmax(
tf.matmul(tf_valid_dataset, weights) + biases)
test_prediction = tf.nn.softmax(tf.matmul(tf_test_dataset, weights) + biases)
In [110]:
with tf.Session(graph=graph) as session:
tf.global_variables_initializer().run()
print("Initialized")
for step in range(num_steps):
# Pick an offset within the training data, which has been randomized.
# Note: we could use better randomization across epochs.
offset = (step * batch_size) % (train_labels.shape[0] - batch_size)
# Generate a minibatch.
batch_data = train_dataset[offset:(offset + batch_size), :]
batch_labels = train_labels[offset:(offset + batch_size), :]
# Prepare a dictionary telling the session where to feed the minibatch.
# The key of the dictionary is the placeholder node of the graph to be fed,
# and the value is the numpy array to feed to it.
feed_dict = {tf_train_dataset : batch_data, tf_train_labels : batch_labels}
_, l, predictions = session.run(
[optimizer, loss, train_prediction], feed_dict=feed_dict)
if (step % 50 == 0):
print("Minibatch loss at step %d: %f" % (step, l))
print("Minibatch accuracy: %.1f%%" % accuracy(predictions, batch_labels))
print("Validation accuracy: %.1f%%" % accuracy(
valid_prediction.eval(), valid_labels))
print("Test accuracy: %.1f%%" % accuracy(test_prediction.eval(), test_labels))
In [111]:
# deeper network
graph = tf.Graph()
with graph.as_default():
# Input data. For the training data, we use a placeholder that will be fed
# at run time with a training minibatch.
tf_train_dataset = tf.placeholder(tf.float32,
shape=(batch_size, image_size * image_size))
tf_train_labels = tf.placeholder(tf.float32, shape=(batch_size, num_labels))
tf_valid_dataset = tf.constant(valid_dataset)
tf_test_dataset = tf.constant(test_dataset)
# Why will weigth1 be of the size 784*h_nodes
# Variables.
weights1 = tf.Variable(
tf.truncated_normal([image_size * image_size, h_nodes]))
biases1 = tf.Variable(tf.zeros([h_nodes]))
weights2 = tf.Variable(
tf.truncated_normal([h_nodes, num_labels]))
biases2 = tf.Variable(tf.zeros([num_labels]))
# Training computation.
logits1 = tf.matmul(tf_train_dataset, weights1) + biases1
# now send these logits to relu
relu_output = tf.nn.relu(logits1)
# introduce dropout to outputs from the relu layer
keep_prob = 0.5
relu_output = tf.nn.dropout(relu_output,keep_prob)
final_logits = tf.matmul(relu_output,weights2) + biases2
loss = tf.reduce_mean(
tf.nn.softmax_cross_entropy_with_logits(labels=tf_train_labels, logits=final_logits))
# now add regularization to it
regularization = tf.nn.l2_loss(weights1) + tf.nn.l2_loss(weights2)
loss = tf.reduce_mean(loss + beta*regularization)
# Optimizer.
optimizer = tf.train.GradientDescentOptimizer(0.5).minimize(loss)
# Predictions for the training, validation, and test data.
train_prediction = tf.nn.softmax(final_logits)
valid_prediction = tf.nn.softmax(tf.matmul(tf.nn.relu(
tf.matmul(tf_valid_dataset, weights1) + biases1),weights2)+biases2)
test_prediction = tf.nn.softmax(tf.matmul(tf.nn.relu(
tf.matmul(tf_test_dataset, weights1) + biases1),weights2)+biases2)
In [112]:
with tf.Session(graph=graph) as session:
tf.global_variables_initializer().run()
print("Initialized")
for step in range(num_steps):
# Pick an offset within the training data, which has been randomized.
# Note: we could use better randomization across epochs.
offset = (step * batch_size) % (train_labels.shape[0] - batch_size)
# Generate a minibatch.
batch_data = train_dataset[offset:(offset + batch_size), :]
batch_labels = train_labels[offset:(offset + batch_size), :]
# Prepare a dictionary telling the session where to feed the minibatch.
# The key of the dictionary is the placeholder node of the graph to be fed,
# and the value is the numpy array to feed to it.
feed_dict = {tf_train_dataset : batch_data, tf_train_labels : batch_labels}
_, l, predictions = session.run(
[optimizer, loss, train_prediction], feed_dict=feed_dict)
if (step % 50 == 0):
print("Minibatch loss at step %d: %f" % (step, l))
print("Minibatch accuracy: %.1f%%" % accuracy(predictions, batch_labels))
print("Validation accuracy: %.1f%%" % accuracy(
valid_prediction.eval(), valid_labels))
print("Test accuracy: %.1f%%" % accuracy(test_prediction.eval(), test_labels))
We conclude that an MLP is not able to perform well on this task
In [ ]: