def test_generate_image_validation_split():
os.listdir = mock_list_image_dir
data_set_struct = make_basic_data_set_image_config()
data_set_config = DatasetConfig(data_set_struct)
train_struct = test_training_config.make_basic_config()
train_struct['validation'] = {
"algorithm": "randomFraction",
"arguments": {
"seed": 1234,
"fraction": 0.3333333
}
}
train_config = TrainingConfig('', train_struct)
dm = jbfs.DataManager({})
train_list, val_list = jb_data.generate_image_list('data_root',
data_set_config,
train_config, dm)
assert len(train_list) == 8
assert len(val_list) == 4
# NOTE: Another fragile secret we know is the order from the validation is is reversed
assert_correct_list(train_list, [1, 2, 4, 5], [1, 2, 3, 4])
assert_correct_list(val_list, [3, 0], [5, 0])
def test_generate_image_sample_fraction():
# If we pass in sampling count we should just get those
# We know how the internal randomizer works. We know it uses random.sample on both
# sets in order. This is a secret and fragile to this test.
# With a seed of 1234 and two pulls of sampling with a count of 2, it pulls [3,0] and [0,5]
os.listdir = mock_list_image_dir
data_set_struct = make_basic_data_set_image_config()
data_set_struct.update(
make_sample_stanza("randomFraction", {
'seed': 1234,
'fraction': 0.3333333333
}))
data_set_config = DatasetConfig(data_set_struct)
dm = jbfs.DataManager({})
train_list, val_list = jb_data.generate_image_list('data_root',
data_set_config, None,
dm)
assert len(train_list) == 4
assert len(val_list) == 0
# Make sure they are in this order
assert_correct_list(train_list, [3, 0], [0, 5])
def test_data_manager():
data_set_config = {}
dm = jbfs.DataManager(data_set_config)
assert dm.version_path == Path()
dm = jbfs.DataManager(data_set_config, version='009')
assert dm.version_path == Path('009')
data_set_config['dataSetPath'] = 'testDataSet'
dm = jbfs.DataManager(data_set_config)
assert dm.version_path == Path('testDataSet')
dm = jbfs.DataManager(data_set_config, version='009')
assert dm.version_path == Path('testDataSet') / '009'
category = 'cars'
assert dm.get_directory_path(
category) == Path('testDataSet') / '009' / 'cars'
image = 'miata.png'
assert dm.get_file_path(
category, image) == Path('testDataSet') / '009' / 'cars' / 'miata.png'
data_set_config['imageData'] = {}
data_set_config['imageData']['properties'] = {}
data_set_config['imageData']['properties']['dimensions'] = '64,64'
data_set_config['imageData']['properties']['colorspace'] = 'gray'
dm = jbfs.DataManager(data_set_config, version='009')
assert dm.check_cache('./test', category, image) == ('miata.png', False)
temp_dir = Path.cwd(
) / 'test' / 'cache' / 'testDataSet' / '009' / '64x64_gray' / 'cars'
temp_image = temp_dir / 'miata.png'
if not temp_dir.exists():
temp_dir.mkdir(parents=True)
temp_image.touch()
import shutil
shutil.rmtree(Path.cwd() / 'test' / 'cache')
def test_generate_image_list():
# Just replace listdir
os.listdir = mock_list_image_dir
data_set_struct = make_basic_data_set_image_config()
data_set_config = DatasetConfig(data_set_struct)
dm = jbfs.DataManager({})
train_list, val_list = jb_data.generate_image_list('data_root',
data_set_config, None,
dm)
assert len(train_list) == 12
assert len(val_list) == 0
assert_correct_list(train_list, range(6), range(6))
def setup(self):
# Construct helper objects
self.data_manager = jbfs.DataManager(self.data_set_config.config, self.data_version)
if juneberry.TENSORBOARD_ROOT:
self.tb_mgr = jbtb.TensorBoardManager(juneberry.TENSORBOARD_ROOT, self.model_manager)
pyt_utils.set_seeds(self.training_config.seed)
self.setup_hardware()
self.setup_data_loaders()
self.setup_model()
self.loss_function = pyt_utils.make_criterion(self.pytorch_options, self.binary)
self.optimizer = pyt_utils.make_optimizer(self.pytorch_options, self.model)
self.lr_scheduler = pyt_utils.make_lr_scheduler(self.pytorch_options, self.optimizer)
self.accuracy_function = pyt_utils.make_accuracy(self.pytorch_options, self.binary)
self.setup_acceptance_checker()
self.num_batches = len(self.training_iterable)
self.history = {'loss': [], 'accuracy': [], 'val_loss': [], 'val_accuracy': [], 'epoch_duration': [], 'lr': []}