2martens/masterthesis
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Optimized check for empty labels

Browse Source 
Signed-off-by: Jim Martens <github@2martens.de>
This commit is contained in:
Jim Martens
2019-04-28 18:38:27 +02:00
parent 2276d359d6
commit ee74480dd7
1 changed files with 4 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
src/twomartens/masterthesis/data.py
View File

@ -258,8 +258,6 @@ def load_scenenet_val(photo_paths: Sequence[Sequence[str]],
traj_image_paths, traj_instances = trajectory
for image_path, frame_instances in zip(traj_image_paths, traj_instances):
labels = []
if not frame_instances: # skip frames without instances
continue
for instance in frame_instances:
bbox = instance['bbox']
labels.append((
@ -269,6 +267,9 @@ def load_scenenet_val(photo_paths: Sequence[Sequence[str]],
bbox[2],
bbox[3]
))
if not labels:
continue
final_image_paths.append(image_path)
final_labels.append(labels)
@ -276,7 +277,7 @@ def load_scenenet_val(photo_paths: Sequence[Sequence[str]],
length_dataset = len(final_image_paths)
path_dataset = tf.data.Dataset.from_tensor_slices(final_image_paths)
label_dataset = tf.data.Dataset.from_sparse_tensor_slices(final_labels)
label_dataset = tf.data.Dataset.from_tensor_slices(final_labels)
dataset = tf.data.Dataset.zip((path_dataset, label_dataset))
dataset = dataset.apply(tf.data.experimental.shuffle_and_repeat(buffer_size=length_dataset, count=num_epochs))
dataset = dataset.batch(batch_size=batch_size)