def test_with(sess, solver, path, name, time_steps=26, batch_size=1):
# Load test instances
print("{timestamp}\t{memory}\tLoading test {name} instances ...".format(
timestamp=timestamp(), memory=memory_usage(), name=name))
test_generator = instance_loader.InstanceLoader(path)
test_loss = 0.0
test_accuracy = 0.0
test_avg_pred = 0.0
test_batches = 0
# Run with the test instances
print("{timestamp}\t{memory}\t{name} TEST SET BEGIN".format(
timestamp=timestamp(), memory=memory_usage(), name=name))
for b, batch in enumerate(test_generator.get_batches(batch_size)):
l, a, p = run_and_log_batch(sess,
solver,
name,
b,
batch,
time_steps,
train=False)
test_loss += l
test_accuracy += a
test_avg_pred += p
test_batches += 1
#end for
# Summarize results and print test summary
test_loss /= test_batches
test_accuracy /= test_batches
test_avg_pred /= test_batches
print(
"{timestamp}\t{memory}\t{name} TEST SET END Mean loss: {loss:.4f} Mean Accuracy = {accuracy} Mean prediction {avg_pred:.4f}"
.format(loss=test_loss,
accuracy=test_accuracy,
avg_pred=test_avg_pred,
timestamp=timestamp(),
memory=memory_usage(),
name=name))
time_steps = 26
batch_size = 128
batches_per_epoch = 128
early_stopping_window = [0 for _ in range(3)]
early_stopping_threshold = 0.85
# Build model
print("{timestamp}\t{memory}\tBuilding model ...".format(
timestamp=timestamp(), memory=memory_usage()))
solver = build_neurosat(d)
# Create batch loader
print("{timestamp}\t{memory}\tLoading instances ...".format(
timestamp=timestamp(), memory=memory_usage()))
generator = instance_loader.InstanceLoader("./instances")
# If you want to use the entire dataset on each epoch, use:
# batches_per_epoch = len(generator.filenames) // batch_size
test_generator = instance_loader.InstanceLoader("./test-instances")
# Create model saver
saver = tf.train.Saver()
# Disallow GPU use
config = tf.ConfigProto(
#device_count = {"GPU":0},
gpu_options=tf.GPUOptions(allow_growth=True), )
with tf.Session(config=config) as sess:
# Initialize global variables
time_steps = 32
batch_size = 4
batches_per_epoch = 128
early_stopping_window = [0 for _ in range(3)]
early_stopping_threshold = 0.85
# Build model
print("{timestamp}\t{memory}\tBuilding model ...".format(
timestamp=timestamp(), memory=memory_usage()))
solver = build_neurosat(d)
# Create batch loader
print("{timestamp}\t{memory}\tLoading instances ...".format(
timestamp=timestamp(), memory=memory_usage()))
generator = instance_loader.InstanceLoader("../adversarial-training-cnf")
# Create model saver
saver = tf.train.Saver()
# Disallow GPU use
config = tf.ConfigProto(device_count={"GPU": 0})
with tf.Session(config=config) as sess:
# Initialize global variables
print(
"{timestamp}\t{memory}\tInitializing global variables ... ".format(
timestamp=timestamp(), memory=memory_usage()))
sess.run(tf.global_variables_initializer())
if os.path.exists("./tmp-64/neurosat.ckpt"):
with tf.Session() as sess:
# Initialize global variables
print(
"{timestamp}\t{memory}\tInitializing global variables ... ".format(
timestamp=timestamp(), memory=memory_usage()))
sess.run(tf.global_variables_initializer())
# Restore saved weights
print("{timestamp}\t{memory}\tRestoring saved model ... ".format(
timestamp=timestamp(), memory=memory_usage()))
saver.restore(sess, "./tmp backup/neurosat.ckpt")
# Define loader and get a batch with size 1 (one instance)
loader = instance_loader.InstanceLoader('test-instances/sat/')
batch = list(itertools.islice(loader.get_batches(1), 1))[0]
time_steps = 100
votes = np.zeros((time_steps, 2 * batch.n[0]))
for t in range(0, time_steps):
votes[t, :] = sess.run(solver["votes"],
feed_dict={
solver["gnn"].time_steps:
t,
solver["gnn"].matrix_placeholders["M"]:
batch.get_dense_matrix(),
solver["instance_SAT"]:
