9f31fde984bac09cfcaf03639f7e86eb09e1f5d1,src/detection/rv/commands/predict_on_chips.py,,_predict_on_chips,#Any#Any#Any#Any#Any#,66
Before Change
vis_util.visualize_boxes_and_labels_on_image_array(
image_np,
np.squeeze(boxes),
np.squeeze(classes).astype(np.int32),
np.squeeze(scores),
category_index,
use_normalized_coordinates=True,
After Change
norm_boxlist = compute_prediction(
image_np, detection_graph, sess)
boxlist = scale(norm_boxlist, width, height)
if predictions_debug_dir is not None:
vis_util.visualize_boxes_and_labels_on_image_array(
image_np,
boxlist.get(),
boxlist.get_field("classes"),
boxlist.get_field("scores"),
category_index,
use_normalized_coordinates=False,
line_thickness=line_thickness)
debug_image_path = \
join(predictions_debug_dir, basename(image_path))
imsave(debug_image_path, image_np)
filename = basename(image_path)
filtered_boxlist = filter_scores_greater_than(
boxlist, min_score_threshold)
predictions[filename] = {
"boxes": filtered_boxlist.get().tolist(),
"scores": filtered_boxlist.get_field("scores").tolist(),
"classes": filtered_boxlist.get_field("classes").tolist()
}
In pattern: SUPERPATTERN
Frequency: 4
Non-data size: 4
Instances
Project Name: azavea/raster-vision
Commit Name: 9f31fde984bac09cfcaf03639f7e86eb09e1f5d1
Time: 2017-10-06
Author: lewfish@gmail.com
File Name: src/detection/rv/commands/predict_on_chips.py
Class Name:
Method Name: _predict_on_chips
Project Name: catalyst-team/catalyst
Commit Name: 1ef3ad90a3423ed15ca41e0ea4e81012ebe84a9f
Time: 2020-08-11
Author: scitator@gmail.com
File Name: catalyst/data/scripts/project_embeddings.py
Class Name:
Method Name: main
Project Name: mil-tokyo/webdnn
Commit Name: 892011b0b264d026d2fdb439580d82de393b5270
Time: 2017-04-21
Author: hidaka@mi.t.u-tokyo.ac.jp
File Name: example/convert_resnet/convert_resnet.py
Class Name:
Method Name: main
Project Name: instacart/lore
Commit Name: 8d8d007fa2ffdf2a7f8f0a5ea596db84f942339e
Time: 2017-12-12
Author: montanalow@users.noreply.github.com
File Name: lore/encoders.py
Class Name: Token
Method Name: transform