def detr_resnet101_panoptic(pretrained=False,
num_classes=91,
threshold=0.85,
return_postprocessor=False):
"""
DETR-DC5 R101 with 6 encoder and 6 decoder layers.
Achieves 45.1 PQ on COCO val5k.
threshold is the minimum confidence required for keeping segments in the prediction
"""
model = _make_detr("resnet101",
dilation=False,
num_classes=num_classes,
mask=True)
is_thing_map = {i: i <= 90 for i in range(250)}
if pretrained:
checkpoint = torch.hub.load_state_dict_from_url(
url=
"https://dl.fbaipublicfiles.com/detr/detr-r101-panoptic-40021d53.pth",
map_location="cpu",
check_hash=True,
)
model.load_state_dict(checkpoint["model"])
if return_postprocessor:
return model, PostProcessPanoptic(is_thing_map, threshold=threshold)
return model
def detr_resnet50_dc5_panoptic(pretrained=False,
num_classes=91,
threshold=0.85,
return_postprocessor=False):
"""
DETR-DC5 R50 with 6 encoder and 6 decoder layers.
The last block of ResNet-50 has dilation to increase
output resolution.
Achieves 44.6 on COCO val5k.
threshold is the minimum confidence required for keeping segments in the prediction
"""
model = _make_detr("resnet50",
dilation=True,
num_classes=num_classes,
mask=True)
is_thing_map = {i: i <= 90 for i in range(250)}
if pretrained:
checkpoint = torch.hub.load_state_dict_from_url(
url=
"https://dl.fbaipublicfiles.com/detr/detr-r50-dc5-panoptic-da08f1b1.pth",
map_location="cpu",
check_hash=True,
)
model.load_state_dict(checkpoint["model"])
if return_postprocessor:
return model, PostProcessPanoptic(is_thing_map, threshold=threshold)
return model
def detr_resnet50_dc5_panoptic(pretrained=False,
num_classes=250,
threshold=0.85,
return_postprocessor=False):
"""
DETR-DC5 R50 with 6 encoder and 6 decoder layers.
The last block of ResNet-50 has dilation to increase
output resolution.
Achieves 44.6 on COCO val5k.
threshold is the minimum confidence required for keeping segments in the prediction
"""
model = _make_detr('resnet50',
dilation=True,
num_classes=num_classes,
mask=True)
is_thing_map = {i: (i <= 90) for i in range(250)}
if pretrained:
checkpoint = paddle.load(
'https://dl.fbaipublicfiles.com/detr/detr-r50-dc5-panoptic-da08f1b1.pdiparams'
)
model.load_state_dict(checkpoint['model'])
if return_postprocessor:
return model, PostProcessPanoptic(is_thing_map, threshold=threshold)
return model
def build(args):
num_classes = 20 if args.dataset_file != 'coco' else 91
if args.dataset_file == "coco_panoptic":
num_classes = 250
device = torch.device(args.device)
backbone = build_backbone(args)
transformer = build_transformer(args)
model = DETR(
args,
backbone,
transformer,
num_classes=num_classes,
num_queries=args.num_queries,
aux_loss=args.aux_loss,
)
if args.masks:
model = DETRsegm(model, freeze_detr=(args.frozen_weights is not None))
matcher = build_matcher(args)
weight_dict = {'loss_ce': 1, 'loss_bbox': args.bbox_loss_coef}
weight_dict['loss_giou'] = args.giou_loss_coef
if args.masks:
weight_dict["loss_mask"] = args.mask_loss_coef
weight_dict["loss_dice"] = args.dice_loss_coef
# TODO this is a hack
if args.aux_loss:
aux_weight_dict = {}
for i in range(args.dec_layers - 1):
aux_weight_dict.update(
{k + f'_{i}': v
for k, v in weight_dict.items()})
weight_dict.update(aux_weight_dict)
losses = ['labels', 'boxes', 'cardinality']
if args.masks:
losses += ["masks"]
criterion = SetCriterion(num_classes,
matcher=matcher,
weight_dict=weight_dict,
eos_coef=args.eos_coef,
losses=losses)
criterion.to(device)
postprocessors = {'bbox': PostProcess()}
if args.masks:
postprocessors['segm'] = PostProcessSegm()
if args.dataset_file == "coco_panoptic":
is_thing_map = {i: i <= 90 for i in range(201)}
postprocessors["panoptic"] = PostProcessPanoptic(is_thing_map,
threshold=0.85)
return model, criterion, postprocessors
def detr_resnet50_panoptic(pretrained=False, num_classes=250, threshold=\
0.85, return_postprocessor=False):
"""
DETR R50 with 6 encoder and 6 decoder layers.
Achieves 43.4 PQ on COCO val5k.
threshold is the minimum confidence required for keeping segments in the prediction
"""
model = _make_detr('resnet50',
dilation=False,
num_classes=num_classes,
mask=True)
is_thing_map = {i: (i <= 90) for i in range(250)}
if pretrained:
checkpoint = paddle.load(
'https://dl.fbaipublicfiles.com/detr/detr-r50-panoptic-00ce5173.pdiparams'
)
model.load_state_dict(checkpoint['model'])
if return_postprocessor:
return model, PostProcessPanoptic(is_thing_map, threshold=threshold)
return model
def build(args):
# the `num_classes` naming here is somewhat misleading.
# it indeed corresponds to `max_obj_id + 1`, where max_obj_id
# is the maximum id for a class in your dataset. For example,
# COCO has a max_obj_id of 90, so we pass `num_classes` to be 91.
# As another example, for a dataset that has a single class with id 1,
# you should pass `num_classes` to be 2 (max_obj_id + 1).
# For more details on this, check the following discussion
# https://github.com/facebookresearch/detr/issues/108#issuecomment-650269223
num_classes = 20 if args.dataset_file != 'coco' else 91
if args.dataset_file == "coco_panoptic":
# for panoptic, we just add a num_classes that is large enough to hold
# max_obj_id + 1, but the exact value doesn't really matter
num_classes = 250
device = torch.device(args.device)
backbone = build_backbone(args)
if int(os.environ.get("cross_transformer", 0)):
transformer = build_cross_transformer(args)
elif int(os.environ.get("sparse_transformer", 0)):
transformer = build_sparse_transformer(args)
elif int(os.environ.get("linear_transformer", 0)):
transformer = build_linear_transformer(args)
else:
transformer = build_transformer(args)
model = DETR(
backbone,
transformer,
num_classes=num_classes,
num_queries=args.num_queries,
aux_loss=args.aux_loss,
)
if args.masks:
model = DETRsegm(model, freeze_detr=(args.frozen_weights is not None))
matcher = build_matcher(args)
weight_dict = {'loss_ce': 1, 'loss_bbox': args.bbox_loss_coef}
weight_dict['loss_giou'] = args.giou_loss_coef
if args.masks:
weight_dict["loss_mask"] = args.mask_loss_coef
weight_dict["loss_dice"] = args.dice_loss_coef
# TODO this is a hack
if args.aux_loss:
aux_weight_dict = {}
for i in range(args.dec_layers - 1):
aux_weight_dict.update(
{k + f'_{i}': v
for k, v in weight_dict.items()})
weight_dict.update(aux_weight_dict)
losses = ['labels', 'boxes', 'cardinality']
if args.masks:
losses += ["masks"]
criterion = SetCriterion(num_classes,
matcher=matcher,
weight_dict=weight_dict,
eos_coef=args.eos_coef,
losses=losses)
criterion.to(device)
postprocessors = {'bbox': PostProcess()}
if args.masks:
postprocessors['segm'] = PostProcessSegm()
if args.dataset_file == "coco_panoptic":
is_thing_map = {i: i <= 90 for i in range(201)}
postprocessors["panoptic"] = PostProcessPanoptic(is_thing_map,
threshold=0.85)
return model, criterion, postprocessors
