def selfie2anime():
img_id = os.environ['id']
result_id = os.environ['result']
parser = get_parser()
args = parser.parse_args("--phase test".split())
with tf.Session(config=tf.ConfigProto(allow_soft_placement=True)) as sess:
#sess.reuse_variables()
gan = UGATIT(sess, args)
# build graph
gan.build_model()
# download target img
download_path = os.path.join(img_path, img_id)
download_image(images_bucket, img_id, dest=download_path)
dataset_tool.create_from_images(record_path, img_path, True)
# os.remove(del_record)
img = gan.infer(download_path)
image_url = upload_image(img, result_id)
return download_path, img
def test_default_argment(self):
parser = main.get_parser().parse_args()
self.assertEqual(parser.epoch, 100)
self.assertEqual(parser.learning_rate, 0.0001)
self.assertEqual(parser.train_rate, 0.8)
self.assertEqual(parser.batch_size, 20)
self.assertEqual(parser.l2, 0.05)
def main():
name = sys.argv[1]
t = sys.argv[2]
value = float(sys.argv[3])
is_int = bool(sys.argv[4])
training_data, testing_data, validation_data = load_data(
"data4students.mat")
#if is_int:
# rand = np.random.random_integers(low, high, samples)
#else:
print("Testing param values for {}={}".format(name, value))
#if is_int:
# params_in = [str(np.random.randint(2000)) for i in range(value)]
#else:
# params_in = [str(value)]
parser = get_parser()
params = parser.parse_args([
"--timestamp",
str(t), "--lr_scheduler", name, "--decay_rate",
str(value)
])
#params = parser.parse_args(["--timestamp", str(t), "--{}".format(name), str(value)])
train_and_report(training_data, validation_data, params)
def main():
if len(sys.argv) != 2:
print("Please specify a path to a .mat file containing the testing data")
return
data = load_data(sys.argv[1])
parser = get_parser()
params = parser.parse_args([])
model = Model(None, None, params)
model.build()
model.model.load_weights('trained_model.h5')
predictions = test_network(model.model, data.data)
print(predictions)
correct = 0
for i, datum in enumerate(data.targets):
datum = list(datum).index(1)
print('Prediction for image ' + str(i) + ': ' + str(predictions[i]) + ', expected ' + str(datum))
correct += 1 if predictions[i] == datum else 0
print('Overall CR: ' + str(correct / len(predictions) * 100) + '%')
def get_parameter():
parser = entry.get_parser()
parser.add_argument('--old', type=str, default='')
parser.add_argument('--new', type=str, default='')
parser.add_argument('--mapping_from', '--mf', type=str, default='')
parser.add_argument('--mapping_to', '--mt', type=str, default='')
parser.add_argument('--verbose_list', default='ratio,sep', type=str)
args = parser.parse_args()
if isinstance(args.verbose_list, str):
args.verbose_list = [x.strip() for x in args.verbose_list.split(',')]
if isinstance(args.keyword, str):
args.keyword = [x.strip() for x in args.keyword.split(',')]
return args
def test_command_sqs():
parser = get_parser()
args = parser.parse_args('sqs --queues queue1 queue2'.split())
assert args.func == start_sqs
assert args.queues == ['queue1', 'queue2']
def test_command_sqs():
parser = get_parser()
args = parser.parse_args('es --ip ip_add --domain foo'.split())
assert args.func == start_es
assert args.ip == 'ip_add'
assert args.domain == 'foo'
