Should these lines refer to the training masks, rather than the validation masks?
|
segmentation_sourcefile = 'validation-annotations-object-segmentation.csv' |
|
segmentation_folder = 'annotations/validation_masks/' |
Here is a larger snippet:
|
if subset == 'train' and args.version != 'challenge_2019': |
|
category_sourcefile = 'class-descriptions-boxable.csv' |
|
image_sourcefile = 'train-images-boxable-with-rotation.csv' |
|
if args.version == 'v6': |
|
annotation_sourcefile = 'oidv6-train-annotations-bbox.csv' |
|
else: |
|
annotation_sourcefile = 'train-annotations-bbox.csv' |
|
image_label_sourcefile = 'train-annotations-human-imagelabels-boxable.csv' |
|
image_size_sourcefile = 'train_sizes-00000-of-00001.csv' |
|
segmentation_sourcefile = 'validation-annotations-object-segmentation.csv' |
|
segmentation_folder = 'annotations/validation_masks/' |
Should these lines refer to the training masks, rather than the validation masks?
openimages2coco/convert_annotations.py
Lines 53 to 54 in 8991d9b
Here is a larger snippet:
openimages2coco/convert_annotations.py
Lines 44 to 54 in 8991d9b