def init_queries(self, query_file, query_config, prefix=''):
queries = read_query_list(
Path(self.base_path, query_file), prefix=prefix)
Dataset = get_dataset(query_config.get('name', self.dataset_name))
query_config = {
**query_config, 'image_names': [q.name for q in queries]}
query_dataset = Dataset(**query_config)
return queries, query_dataset
def init_queries(self, query_file, query_config, prefix=""):
queries = read_query_list(Path(self.base_path, query_file),
prefix=prefix)
Dataset = get_dataset(query_config.get("name", self.dataset_name))
query_config = {
**query_config, "image_names": [q.name for q in queries]
}
query_dataset = Dataset(**query_config)
# load GPS data for queries
query_gps = Traverse(self.dataset_name, self.config["queries"],
self.gps.experiment_name)
return queries, query_dataset, query_gps
def _init_graph(config, with_dataset=False):
set_seed(config.get('seed', int.from_bytes(os.urandom(4), byteorder='big')))
n_gpus = len(os.environ['CUDA_VISIBLE_DEVICES'].split(','))
logging.info('Number of GPUs detected: {}'.format(n_gpus))
dataset = get_dataset(config['data']['name'])(**config['data'])
model = get_model(config['model']['name'])(
data=dataset.get_tf_datasets(), n_gpus=n_gpus, **config['model'])
model.__enter__()
if with_dataset:
yield model, dataset
else:
yield model
model.__exit__()
tf.reset_default_graph()
def export_for_sfm(data_config, exper_config, export_name):
export_dir = Path(EXPER_PATH, 'exports', export_name)
export_dir.mkdir(exist_ok=True)
dataset = get_dataset(data_config['name'])(**data_config)
data_iter = dataset.get_test_set()
for data in tqdm(data_iter):
predictions = exper_config['predictor'](
data['image'], data['name'], **exper_config)
# Scale the keypoints to the original image size
# and convert to Colmap convention (origin = corner of upper left pix)
scale = ((np.array(data['original_size'][:2]) - 1)
/ (np.array(data['image'].shape[:2]) - 1))
export = {
'keypoints': scale[::-1] * predictions['keypoints'] + 0.5,
'scores': predictions['scores'],
'descriptors': predictions['descriptors'],
'image_size': data['image'].shape[:2][::-1]
}
name = data['name'].decode('utf-8')
Path(export_dir, Path(name).parent).mkdir(parents=True, exist_ok=True)
np.savez(Path(export_dir, f'{name}.npz'), **export)
with open(Path(EXPER_PATH, exper_name, 'config.yaml'), 'r') as f:
config['model'] = tools.dict_update(
yaml.load(f)['model'], config.get('model', {}))
checkpoint_path = Path(EXPER_PATH, exper_name)
if config.get('weights', None):
checkpoint_path = Path(checkpoint_path, config['weights'])
else:
if config.get('weights', None):
checkpoint_path = Path(DATA_PATH, 'weights', config['weights'])
else:
checkpoint_path = None
logging.info('No weights provided.')
logging.info(f'Starting export with configuration:\n{pformat(config)}')
with get_model(config['model']['name'])(data_shape={
'image': [None, None, None, config['model']['image_channels']]
},
**config['model']) as net:
if checkpoint_path is not None:
net.load(str(checkpoint_path))
dataset = get_dataset(config['data']['name'])(**config['data'])
test_set = dataset.get_test_set()
for data in tqdm(test_set):
predictions = net.predict(data, keys=keys)
predictions['input_shape'] = data['image'].shape
name = data['name'].decode('utf-8')
Path(base_dir,
Path(name).parent).mkdir(parents=True, exist_ok=True)
np.savez(Path(base_dir, '{}.npz'.format(name)), **predictions)
