def __init__(self, cfg, in_channels):
super(KeypointRCNNFeatureExtractor, self).__init__()
use_gn = cfg.MODEL.ROI_KEYPOINT_HEAD.USE_GN
use_contextual_pooler = False
if use_contextual_pooler:
pooler = make_contextual_pooler(cfg, 'ROI_KEYPOINT_HEAD')
else:
pooler = make_pooler(cfg, 'ROI_KEYPOINT_HEAD')
self.pooler = pooler
input_features = in_channels
layers = cfg.MODEL.ROI_KEYPOINT_HEAD.CONV_LAYERS
resolution = cfg.MODEL.ROI_KEYPOINT_HEAD.POOLER_RESOLUTION
next_feature = input_features
self.blocks = []
for layer_idx, layer_features in enumerate(layers, 1):
layer_name = "kp_fcn{}".format(layer_idx)
module = make_conv3x3(
next_feature, layer_features,
dilation=1, stride=1, use_gn=use_gn, use_relu=True, kaiming_init=True
)
self.add_module(layer_name, module)
next_feature = layer_features
self.blocks.append(layer_name)
self.out_channels = layer_features
if cfg.MODEL.ROI_KEYPOINT_HEAD.ATTENTION_ON:
self.regional_attention = RegionalAttention(cfg, in_channels, self.pooler, resolution)
else:
self.regional_attention = None
def __init__(self, cfg, in_channels):
"""
Arguments:
num_classes (int): number of output classes
input_size (int): number of channels of the input once it's flattened
representation_size (int): size of the intermediate representation
"""
super(MaskRCNNFPNFeatureExtractor, self).__init__()
use_contextual_pooler = False
if use_contextual_pooler:
pooler = make_contextual_pooler(cfg, 'ROI_MASK_HEAD')
else:
pooler = make_pooler(cfg, 'ROI_MASK_HEAD')
input_size = in_channels
self.pooler = pooler
use_gn = cfg.MODEL.ROI_MASK_HEAD.USE_GN
layers = cfg.MODEL.ROI_MASK_HEAD.CONV_LAYERS
dilation = cfg.MODEL.ROI_MASK_HEAD.DILATION
resolution = cfg.MODEL.ROI_MASK_HEAD.POOLER_RESOLUTION
use_dcn = cfg.MODEL.ROI_MASK_HEAD.USE_DCN
dcn_config = {
"with_modulated_dcn": cfg.MODEL.ROI_MASK_HEAD.WITH_MODULATED_DCN,
"deformable_groups": cfg.MODEL.ROI_MASK_HEAD.DEFORMABLE_GROUPS,
}
self.blocks, next_feature = make_blocks(
layers,
input_size,
pattern='mask_fcn',
parent=self,
dilation=dilation,
use_gn=use_gn,
use_dcn=use_dcn,
dcn_config=dcn_config,
)
self.out_channels = next_feature
if cfg.MODEL.ROI_MASK_HEAD.ATTENTION_ON:
self.regional_attention = RegionalAttention(
cfg, in_channels, self.pooler, resolution)
else:
self.regional_attention = None
def __init__(self, cfg, in_channels):
super(FPNXconv1fcFeatureExtractor, self).__init__()
resolution = cfg.MODEL.ROI_BOX_HEAD.POOLER_RESOLUTION
use_contextual_pooler = False
if use_contextual_pooler:
pooler = make_contextual_pooler(cfg, 'ROI_BOX_HEAD')
else:
pooler = make_pooler(cfg, 'ROI_BOX_HEAD')
input_size = in_channels
self.pooler = pooler
use_gn = cfg.MODEL.ROI_BOX_HEAD.USE_GN
dilation = cfg.MODEL.ROI_BOX_HEAD.DILATION
layers = cfg.MODEL.ROI_MASK_HEAD.CONV_LAYERS
use_dcn = cfg.MODEL.ROI_BOX_HEAD.USE_DCN
dcn_config = {
"with_modulated_dcn": cfg.MODEL.ROI_BOX_HEAD.WITH_MODULATED_DCN,
"deformable_groups": cfg.MODEL.ROI_BOX_HEAD.DEFORMABLE_GROUPS,
}
self.xconvs, next_feature = make_blocks(
layers,
input_size,
pattern='xconvs',
parent=self,
dilation=dilation,
use_gn=use_gn,
use_dcn=use_dcn,
dcn_config=dcn_config,
)
input_size = next_feature * resolution**2
representation_size = cfg.MODEL.ROI_BOX_HEAD.MLP_HEAD_DIM
self.fc6 = make_fc(input_size, representation_size, use_gn=False)
self.out_channels = representation_size
if cfg.MODEL.ROI_BOX_HEAD.ATTENTION_ON:
self.regional_attention = RegionalAttention(
cfg, in_channels, self.pooler, resolution)
else:
self.regional_attention = None
def __init__(self, cfg, in_channels):
super(FPN2MLPFeatureExtractor, self).__init__()
resolution = cfg.MODEL.ROI_BOX_HEAD.POOLER_RESOLUTION
use_contextual_pooler = False
if use_contextual_pooler:
pooler = make_contextual_pooler(cfg, 'ROI_BOX_HEAD')
else:
pooler = make_pooler(cfg, 'ROI_BOX_HEAD')
input_size = in_channels * resolution**2
representation_size = cfg.MODEL.ROI_BOX_HEAD.MLP_HEAD_DIM
use_gn = cfg.MODEL.ROI_BOX_HEAD.USE_GN
self.pooler = pooler
self.fc6 = make_fc(input_size, representation_size, use_gn)
self.fc7 = make_fc(representation_size, representation_size, use_gn)
self.out_channels = representation_size
def __init__(self, cfg, in_channels):
super(ResNet50Conv5ROIFeatureExtractor, self).__init__()
use_contextual_pooler = False
if use_contextual_pooler:
pooler = make_contextual_pooler(cfg, 'ROI_BOX_HEAD')
else:
pooler = make_pooler(cfg, 'ROI_BOX_HEAD')
stage = resnet.StageSpec(index=4, block_count=3, return_features=False)
head = resnet.ResNetHead(
block_module=cfg.MODEL.RESNETS.TRANS_FUNC,
stages=(stage, ),
num_groups=cfg.MODEL.RESNETS.NUM_GROUPS,
width_per_group=cfg.MODEL.RESNETS.WIDTH_PER_GROUP,
stride_in_1x1=cfg.MODEL.RESNETS.STRIDE_IN_1X1,
stride_init=None,
res2_out_channels=cfg.MODEL.RESNETS.RES2_OUT_CHANNELS,
dilation=cfg.MODEL.RESNETS.RES5_DILATION)
self.pooler = pooler
self.head = head
self.out_channels = head.out_channels
