def parser_generator(params, mode):
"""Generator function for various dataset parser."""
if params.architecture.parser == 'retinanet_parser':
anchor_params = params.anchor
parser_params = params.retinanet_parser
parser_fn = retinanet_parser.Parser(
output_size=parser_params.output_size,
min_level=anchor_params.min_level,
max_level=anchor_params.max_level,
num_scales=anchor_params.num_scales,
aspect_ratios=anchor_params.aspect_ratios,
anchor_size=anchor_params.anchor_size,
match_threshold=parser_params.match_threshold,
unmatched_threshold=parser_params.unmatched_threshold,
aug_rand_hflip=parser_params.aug_rand_hflip,
aug_scale_min=parser_params.aug_scale_min,
aug_scale_max=parser_params.aug_scale_max,
use_autoaugment=parser_params.use_autoaugment,
autoaugment_policy_name=parser_params.autoaugment_policy_name,
skip_crowd_during_training=parser_params.
skip_crowd_during_training,
max_num_instances=parser_params.max_num_instances,
use_bfloat16=parser_params.use_bfloat16,
mode=mode,
regenerate_source_id=parser_params.regenerate_source_id)
elif params.architecture.parser == 'shapemask_parser':
anchor_params = params.anchor
parser_params = params.shapemask_parser
parser_fn = shapemask_parser.Parser(
output_size=parser_params.output_size,
min_level=anchor_params.min_level,
max_level=anchor_params.max_level,
num_scales=anchor_params.num_scales,
aspect_ratios=anchor_params.aspect_ratios,
anchor_size=anchor_params.anchor_size,
use_category=parser_params.use_category,
outer_box_scale=parser_params.outer_box_scale,
box_jitter_scale=parser_params.box_jitter_scale,
num_sampled_masks=parser_params.num_sampled_masks,
mask_crop_size=parser_params.mask_crop_size,
mask_min_level=parser_params.mask_min_level,
mask_max_level=parser_params.mask_max_level,
upsample_factor=parser_params.upsample_factor,
match_threshold=parser_params.match_threshold,
unmatched_threshold=parser_params.unmatched_threshold,
aug_rand_hflip=parser_params.aug_rand_hflip,
aug_scale_min=parser_params.aug_scale_min,
aug_scale_max=parser_params.aug_scale_max,
skip_crowd_during_training=parser_params.
skip_crowd_during_training,
max_num_instances=parser_params.max_num_instances,
use_bfloat16=parser_params.use_bfloat16,
mask_train_class=parser_params.mask_train_class,
mode=mode)
else:
raise ValueError('Parser %s is not supported.' %
params.architecture.parser)
return parser_fn
def parser_generator(params, mode):
"""Generator function for various dataset parser."""
if params.architecture.parser == 'classification_parser':
parser_params = params.classification_parser
parser_fn = classification_parser.Parser(
output_size=parser_params.output_size,
aug_rand_hflip=parser_params.aug_rand_hflip,
use_bfloat16=params.architecture.use_bfloat16,
mode=mode)
elif params.architecture.parser == 'retinanet_parser':
anchor_params = params.anchor
parser_params = params.retinanet_parser
parser_fn = retinanet_parser.Parser(
output_size=parser_params.output_size,
min_level=params.architecture.min_level,
max_level=params.architecture.max_level,
num_scales=anchor_params.num_scales,
aspect_ratios=anchor_params.aspect_ratios,
anchor_size=anchor_params.anchor_size,
match_threshold=parser_params.match_threshold,
unmatched_threshold=parser_params.unmatched_threshold,
aug_rand_hflip=parser_params.aug_rand_hflip,
aug_scale_min=parser_params.aug_scale_min,
aug_scale_max=parser_params.aug_scale_max,
aug_policy=parser_params.aug_policy,
skip_crowd_during_training=parser_params.
skip_crowd_during_training,
max_num_instances=parser_params.max_num_instances,
use_bfloat16=params.architecture.use_bfloat16,
mode=mode,
regenerate_source_id=parser_params.regenerate_source_id)
elif params.architecture.parser == 'maskrcnn_parser':
anchor_params = params.anchor
parser_params = params.maskrcnn_parser
parser_fn = maskrcnn_parser.Parser(
output_size=parser_params.output_size,
min_level=params.architecture.min_level,
max_level=params.architecture.max_level,
num_scales=anchor_params.num_scales,
aspect_ratios=anchor_params.aspect_ratios,
anchor_size=anchor_params.anchor_size,
rpn_match_threshold=parser_params.rpn_match_threshold,
rpn_unmatched_threshold=parser_params.rpn_unmatched_threshold,
rpn_batch_size_per_im=parser_params.rpn_batch_size_per_im,
rpn_fg_fraction=parser_params.rpn_fg_fraction,
aug_rand_hflip=parser_params.aug_rand_hflip,
aug_scale_min=parser_params.aug_scale_min,
aug_scale_max=parser_params.aug_scale_max,
skip_crowd_during_training=parser_params.
skip_crowd_during_training,
max_num_instances=parser_params.max_num_instances,
include_mask=params.architecture.include_mask,
mask_crop_size=parser_params.mask_crop_size,
use_bfloat16=params.architecture.use_bfloat16,
mode=mode)
elif params.architecture.parser == 'shapemask_parser':
anchor_params = params.anchor
parser_params = params.shapemask_parser
parser_fn = shapemask_parser.Parser(
output_size=parser_params.output_size,
min_level=params.architecture.min_level,
max_level=params.architecture.max_level,
num_scales=anchor_params.num_scales,
aspect_ratios=anchor_params.aspect_ratios,
anchor_size=anchor_params.anchor_size,
use_category=parser_params.use_category,
outer_box_scale=parser_params.outer_box_scale,
box_jitter_scale=parser_params.box_jitter_scale,
num_sampled_masks=parser_params.num_sampled_masks,
mask_crop_size=parser_params.mask_crop_size,
mask_min_level=parser_params.mask_min_level,
mask_max_level=parser_params.mask_max_level,
upsample_factor=parser_params.upsample_factor,
match_threshold=parser_params.match_threshold,
unmatched_threshold=parser_params.unmatched_threshold,
aug_rand_hflip=parser_params.aug_rand_hflip,
aug_scale_min=parser_params.aug_scale_min,
aug_scale_max=parser_params.aug_scale_max,
skip_crowd_during_training=parser_params.
skip_crowd_during_training,
max_num_instances=parser_params.max_num_instances,
use_bfloat16=params.architecture.use_bfloat16,
mask_train_class=parser_params.mask_train_class,
mode=mode)
elif params.architecture.parser == 'segmentation_parser':
parser_params = params.segmentation_parser
parser_fn = segmentation_parser.Parser(
output_size=parser_params.output_size,
resize_eval=parser_params.resize_eval,
ignore_label=parser_params.ignore_label,
aug_rand_hflip=parser_params.aug_rand_hflip,
aug_scale_min=parser_params.aug_scale_min,
aug_scale_max=parser_params.aug_scale_max,
use_bfloat16=params.architecture.use_bfloat16,
mode=mode)
else:
raise ValueError('Parser %s is not supported.' %
params.architecture.parser)
return parser_fn
def parser_generator(params, mode):
"""Generator function for various dataset parser."""
if params.architecture.parser == 'classification_parser':
parser_params = params.classification_parser
parser_fn = classification_parser.Parser(
output_size=parser_params.output_size,
aug_rand_hflip=parser_params.aug_rand_hflip,
use_bfloat16=params.architecture.use_bfloat16,
mode=mode)
elif params.architecture.parser == 'retinanet_parser':
anchor_params = params.anchor
parser_params = params.retinanet_parser
parser_fn = retinanet_parser.Parser(
output_size=parser_params.output_size,
min_level=params.architecture.min_level,
max_level=params.architecture.max_level,
num_scales=anchor_params.num_scales,
aspect_ratios=anchor_params.aspect_ratios,
anchor_size=anchor_params.anchor_size,
match_threshold=parser_params.match_threshold,
unmatched_threshold=parser_params.unmatched_threshold,
aug_rand_hflip=parser_params.aug_rand_hflip,
aug_scale_min=parser_params.aug_scale_min,
aug_scale_max=parser_params.aug_scale_max,
aug_policy=parser_params.aug_policy,
skip_crowd_during_training=parser_params.
skip_crowd_during_training,
max_num_instances=parser_params.max_num_instances,
use_bfloat16=params.architecture.use_bfloat16,
mode=mode,
regenerate_source_id=parser_params.regenerate_source_id)
elif params.architecture.parser == 'maskrcnn_parser':
anchor_params = params.anchor
parser_params = params.maskrcnn_parser
parser_fn = maskrcnn_parser.Parser(
output_size=parser_params.output_size,
min_level=params.architecture.min_level,
max_level=params.architecture.max_level,
num_scales=anchor_params.num_scales,
aspect_ratios=anchor_params.aspect_ratios,
anchor_size=anchor_params.anchor_size,
rpn_match_threshold=parser_params.rpn_match_threshold,
rpn_unmatched_threshold=parser_params.rpn_unmatched_threshold,
rpn_batch_size_per_im=parser_params.rpn_batch_size_per_im,
rpn_fg_fraction=parser_params.rpn_fg_fraction,
aug_rand_hflip=parser_params.aug_rand_hflip,
aug_scale_min=parser_params.aug_scale_min,
aug_scale_max=parser_params.aug_scale_max,
skip_crowd_during_training=parser_params.
skip_crowd_during_training,
max_num_instances=parser_params.max_num_instances,
include_mask=params.architecture.include_mask,
mask_crop_size=parser_params.mask_crop_size,
use_bfloat16=params.architecture.use_bfloat16,
regenerate_source_id=parser_params.regenerate_source_id,
mode=mode)
if mode == ModeKeys.TRAIN and parser_params.copy_paste:
parser_fn = maskrcnn_parser_with_copy_paste.Parser(
output_size=parser_params.output_size,
min_level=params.architecture.min_level,
max_level=params.architecture.max_level,
num_scales=anchor_params.num_scales,
aspect_ratios=anchor_params.aspect_ratios,
anchor_size=anchor_params.anchor_size,
rpn_match_threshold=parser_params.rpn_match_threshold,
rpn_unmatched_threshold=parser_params.rpn_unmatched_threshold,
rpn_batch_size_per_im=parser_params.rpn_batch_size_per_im,
rpn_fg_fraction=parser_params.rpn_fg_fraction,
aug_rand_hflip=parser_params.aug_rand_hflip,
aug_scale_min=parser_params.aug_scale_min,
aug_scale_max=parser_params.aug_scale_max,
skip_crowd_during_training=parser_params.
skip_crowd_during_training,
max_num_instances=parser_params.max_num_instances,
include_mask=params.architecture.include_mask,
mask_crop_size=parser_params.mask_crop_size,
use_bfloat16=params.architecture.use_bfloat16,
mode=mode)
elif params.architecture.parser == 'vild_parser':
arch_params = params.architecture
anchor_params = params.anchor
parser_params = params.vild_parser
parser_fn = vild_parser.Parser(
output_size=parser_params.output_size,
min_level=params.architecture.min_level,
max_level=params.architecture.max_level,
num_scales=anchor_params.num_scales,
aspect_ratios=anchor_params.aspect_ratios,
anchor_size=anchor_params.anchor_size,
rpn_match_threshold=parser_params.rpn_match_threshold,
rpn_unmatched_threshold=parser_params.rpn_unmatched_threshold,
rpn_batch_size_per_im=parser_params.rpn_batch_size_per_im,
rpn_fg_fraction=parser_params.rpn_fg_fraction,
aug_rand_hflip=parser_params.aug_rand_hflip,
aug_scale_min=parser_params.aug_scale_min,
aug_scale_max=parser_params.aug_scale_max,
skip_crowd_during_training=parser_params.
skip_crowd_during_training,
max_num_instances=parser_params.max_num_instances,
include_mask=params.architecture.include_mask,
mask_crop_size=parser_params.mask_crop_size,
use_bfloat16=params.architecture.use_bfloat16,
mode=mode,
# ViLD
visual_feature_distill=mode == ModeKeys.TRAIN and
arch_params.visual_feature_distill in ['vanilla', 'double_branch'],
visual_feature_dim=arch_params.visual_feature_dim,
max_num_rois=arch_params.max_num_rois,
filter_distill_boxes_size=arch_params.filter_distill_boxes_size,
)
elif params.architecture.parser == 'extract_objects_parser':
parser_params = params.maskrcnn_parser
parser_fn = extract_objects_parser.Parser(
output_size=parser_params.output_size,
min_level=params.architecture.min_level,
max_level=params.architecture.max_level,
aug_rand_hflip=parser_params.aug_rand_hflip,
aug_scale_min=parser_params.aug_scale_min,
aug_scale_max=parser_params.aug_scale_max,
skip_crowd_during_training=parser_params.
skip_crowd_during_training,
include_mask=params.architecture.include_mask,
mask_crop_size=parser_params.mask_crop_size)
elif params.architecture.parser == 'shapemask_parser':
anchor_params = params.anchor
parser_params = params.shapemask_parser
parser_fn = shapemask_parser.Parser(
output_size=parser_params.output_size,
min_level=params.architecture.min_level,
max_level=params.architecture.max_level,
num_scales=anchor_params.num_scales,
aspect_ratios=anchor_params.aspect_ratios,
anchor_size=anchor_params.anchor_size,
use_category=parser_params.use_category,
outer_box_scale=parser_params.outer_box_scale,
box_jitter_scale=parser_params.box_jitter_scale,
num_sampled_masks=parser_params.num_sampled_masks,
mask_crop_size=parser_params.mask_crop_size,
mask_min_level=parser_params.mask_min_level,
mask_max_level=parser_params.mask_max_level,
upsample_factor=parser_params.upsample_factor,
match_threshold=parser_params.match_threshold,
unmatched_threshold=parser_params.unmatched_threshold,
aug_rand_hflip=parser_params.aug_rand_hflip,
aug_scale_min=parser_params.aug_scale_min,
aug_scale_max=parser_params.aug_scale_max,
skip_crowd_during_training=parser_params.
skip_crowd_during_training,
max_num_instances=parser_params.max_num_instances,
use_bfloat16=params.architecture.use_bfloat16,
mask_train_class=parser_params.mask_train_class,
mode=mode)
elif params.architecture.parser == 'segmentation_parser':
parser_params = params.segmentation_parser
parser_fn = segmentation_parser.Parser(
output_size=parser_params.output_size,
resize_eval=parser_params.resize_eval,
ignore_label=parser_params.ignore_label,
aug_rand_hflip=parser_params.aug_rand_hflip,
aug_scale_min=parser_params.aug_scale_min,
aug_scale_max=parser_params.aug_scale_max,
aug_policy=parser_params.aug_policy,
use_bfloat16=params.architecture.use_bfloat16,
mode=mode)
else:
raise ValueError('Parser %s is not supported.' %
params.architecture.parser)
return parser_fn
