Python maskrcnn_benchmark.modeling.make_layers.make_conv3x3() Examples

The following are 16 code examples of maskrcnn_benchmark.modeling.make_layers.make_conv3x3(). You can vote up the ones you like or vote down the ones you don't like, and go to the original project or source file by following the links above each example. You may also want to check out all available functions/classes of the module maskrcnn_benchmark.modeling.make_layers , or try the search function .
Example #1
Source File: roi_mask_feature_extractors.py    From Res2Net-maskrcnn with MIT License 5 votes vote down vote up
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__()

        resolution = cfg.MODEL.ROI_MASK_HEAD.POOLER_RESOLUTION
        scales = cfg.MODEL.ROI_MASK_HEAD.POOLER_SCALES
        sampling_ratio = cfg.MODEL.ROI_MASK_HEAD.POOLER_SAMPLING_RATIO
        pooler = Pooler(
            output_size=(resolution, resolution),
            scales=scales,
            sampling_ratio=sampling_ratio,
        )
        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

        next_feature = input_size
        self.blocks = []
        for layer_idx, layer_features in enumerate(layers, 1):
            layer_name = "mask_fcn{}".format(layer_idx)
            module = make_conv3x3(
                next_feature, layer_features,
                dilation=dilation, stride=1, use_gn=use_gn
            )
            self.add_module(layer_name, module)
            next_feature = layer_features
            self.blocks.append(layer_name)
        self.out_channels = layer_features 
Example #2
Source File: roi_mask_feature_extractors.py    From R2CNN.pytorch with MIT License 5 votes vote down vote up
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__()

        resolution = cfg.MODEL.ROI_MASK_HEAD.POOLER_RESOLUTION
        scales = cfg.MODEL.ROI_MASK_HEAD.POOLER_SCALES
        sampling_ratio = cfg.MODEL.ROI_MASK_HEAD.POOLER_SAMPLING_RATIO
        pooler = Pooler(
            output_size=(resolution, resolution),
            scales=scales,
            sampling_ratio=sampling_ratio,
        )
        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

        next_feature = input_size
        self.blocks = []
        for layer_idx, layer_features in enumerate(layers, 1):
            layer_name = "mask_fcn{}".format(layer_idx)
            module = make_conv3x3(
                next_feature, layer_features,
                dilation=dilation, stride=1, use_gn=use_gn
            )
            self.add_module(layer_name, module)
            next_feature = layer_features
            self.blocks.append(layer_name)
        self.out_channels = layer_features 
Example #3
Source File: roi_mask_feature_extractors.py    From Clothing-Detection with GNU General Public License v3.0 5 votes vote down vote up
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__()

        resolution = cfg.MODEL.ROI_MASK_HEAD.POOLER_RESOLUTION
        scales = cfg.MODEL.ROI_MASK_HEAD.POOLER_SCALES
        sampling_ratio = cfg.MODEL.ROI_MASK_HEAD.POOLER_SAMPLING_RATIO
        pooler = Pooler(
            output_size=(resolution, resolution),
            scales=scales,
            sampling_ratio=sampling_ratio,
        )
        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

        next_feature = input_size
        self.blocks = []
        for layer_idx, layer_features in enumerate(layers, 1):
            layer_name = "mask_fcn{}".format(layer_idx)
            module = make_conv3x3(
                next_feature, layer_features,
                dilation=dilation, stride=1, use_gn=use_gn
            )
            self.add_module(layer_name, module)
            next_feature = layer_features
            self.blocks.append(layer_name)
        self.out_channels = layer_features 
Example #4
Source File: roi_mask_feature_extractors.py    From DetNAS with MIT License 5 votes vote down vote up
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__()

        resolution = cfg.MODEL.ROI_MASK_HEAD.POOLER_RESOLUTION
        scales = cfg.MODEL.ROI_MASK_HEAD.POOLER_SCALES
        sampling_ratio = cfg.MODEL.ROI_MASK_HEAD.POOLER_SAMPLING_RATIO
        pooler = Pooler(
            output_size=(resolution, resolution),
            scales=scales,
            sampling_ratio=sampling_ratio,
        )
        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

        next_feature = input_size
        self.blocks = []
        for layer_idx, layer_features in enumerate(layers, 1):
            layer_name = "mask_fcn{}".format(layer_idx)
            module = make_conv3x3(
                next_feature, layer_features,
                dilation=dilation, stride=1, use_gn=use_gn
            )
            self.add_module(layer_name, module)
            next_feature = layer_features
            self.blocks.append(layer_name)
        self.out_channels = layer_features 
Example #5
Source File: roi_mask_feature_extractors.py    From remote_sensing_object_detection_2019 with MIT License 5 votes vote down vote up
def __init__(self, cfg):
        """
        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__()

        resolution = cfg.MODEL.ROI_MASK_HEAD.POOLER_RESOLUTION
        scales = cfg.MODEL.ROI_MASK_HEAD.POOLER_SCALES
        sampling_ratio = cfg.MODEL.ROI_MASK_HEAD.POOLER_SAMPLING_RATIO
        pooler = PyramidRROIAlign(
            output_size=(resolution, resolution),
            scales=scales,
        )
        input_size = cfg.MODEL.BACKBONE.OUT_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

        self.word_margin = cfg.MODEL.ROI_REC_HEAD.BOXES_MARGIN
        self.det_margin = cfg.MODEL.RRPN.GT_BOX_MARGIN

        self.rescale = self.word_margin / self.det_margin

        next_feature = input_size
        self.blocks = []
        for layer_idx, layer_features in enumerate(layers, 1):
            layer_name = "mask_fcn{}".format(layer_idx)
            module = make_conv3x3(next_feature, layer_features, 
                dilation=dilation, stride=1, use_gn=use_gn
            )
            self.add_module(layer_name, module)
            next_feature = layer_features
            self.blocks.append(layer_name) 
Example #6
Source File: roi_rec_feature_extractors.py    From remote_sensing_object_detection_2019 with MIT License 5 votes vote down vote up
def __init__(self, cfg):
        """
        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__()

        resolution = cfg.MODEL.ROI_MASK_HEAD.POOLER_RESOLUTION
        scales = cfg.MODEL.ROI_MASK_HEAD.POOLER_SCALES
        sampling_ratio = cfg.MODEL.ROI_MASK_HEAD.POOLER_SAMPLING_RATIO
        pooler = Pooler(
            output_size=(resolution, resolution),
            scales=scales,
            sampling_ratio=sampling_ratio
        )
        input_size = cfg.MODEL.BACKBONE.OUT_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

        next_feature = input_size
        self.blocks = []
        for layer_idx, layer_features in enumerate(layers, 1):
            layer_name = "mask_fcn{}".format(layer_idx)
            module = make_conv3x3(next_feature, layer_features, 
                dilation=dilation, stride=1, use_gn=use_gn
            )
            self.add_module(layer_name, module)
            next_feature = layer_features
            self.blocks.append(layer_name) 
Example #7
Source File: roi_mask_feature_extractors.py    From remote_sensing_object_detection_2019 with MIT License 5 votes vote down vote up
def __init__(self, cfg):
        """
        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__()

        resolution = cfg.MODEL.ROI_MASK_HEAD.POOLER_RESOLUTION
        scales = cfg.MODEL.ROI_MASK_HEAD.POOLER_SCALES
        sampling_ratio = cfg.MODEL.ROI_MASK_HEAD.POOLER_SAMPLING_RATIO
        pooler = Pooler(
            output_size=(resolution, resolution),
            scales=scales,
            sampling_ratio=sampling_ratio,
        )
        input_size = cfg.MODEL.BACKBONE.OUT_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

        next_feature = input_size
        self.blocks = []
        for layer_idx, layer_features in enumerate(layers, 1):
            layer_name = "mask_fcn{}".format(layer_idx)
            module = make_conv3x3(next_feature, layer_features, 
                dilation=dilation, stride=1, use_gn=use_gn
            )
            self.add_module(layer_name, module)
            next_feature = layer_features
            self.blocks.append(layer_name) 
Example #8
Source File: roi_mask_feature_extractors.py    From sampling-free with MIT License 5 votes vote down vote up
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__()

        resolution = cfg.MODEL.ROI_MASK_HEAD.POOLER_RESOLUTION
        scales = cfg.MODEL.ROI_MASK_HEAD.POOLER_SCALES
        sampling_ratio = cfg.MODEL.ROI_MASK_HEAD.POOLER_SAMPLING_RATIO
        pooler = Pooler(
            output_size=(resolution, resolution),
            scales=scales,
            sampling_ratio=sampling_ratio,
        )
        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

        next_feature = input_size
        self.blocks = []
        for layer_idx, layer_features in enumerate(layers, 1):
            layer_name = "mask_fcn{}".format(layer_idx)
            module = make_conv3x3(
                next_feature, layer_features,
                dilation=dilation, stride=1, use_gn=use_gn
            )
            self.add_module(layer_name, module)
            next_feature = layer_features
            self.blocks.append(layer_name)
        self.out_channels = layer_features 
Example #9
Source File: roi_mask_feature_extractors.py    From maskrcnn-benchmark with MIT License 5 votes vote down vote up
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__()

        resolution = cfg.MODEL.ROI_MASK_HEAD.POOLER_RESOLUTION
        scales = cfg.MODEL.ROI_MASK_HEAD.POOLER_SCALES
        sampling_ratio = cfg.MODEL.ROI_MASK_HEAD.POOLER_SAMPLING_RATIO
        pooler = Pooler(
            output_size=(resolution, resolution),
            scales=scales,
            sampling_ratio=sampling_ratio,
        )
        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

        next_feature = input_size
        self.blocks = []
        for layer_idx, layer_features in enumerate(layers, 1):
            layer_name = "mask_fcn{}".format(layer_idx)
            module = make_conv3x3(
                next_feature, layer_features,
                dilation=dilation, stride=1, use_gn=use_gn
            )
            self.add_module(layer_name, module)
            next_feature = layer_features
            self.blocks.append(layer_name)
        self.out_channels = layer_features 
Example #10
Source File: roi_mask_feature_extractors.py    From training with Apache License 2.0 5 votes vote down vote up
def __init__(self, cfg):
        """
        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__()

        resolution = cfg.MODEL.ROI_MASK_HEAD.POOLER_RESOLUTION
        scales = cfg.MODEL.ROI_MASK_HEAD.POOLER_SCALES
        sampling_ratio = cfg.MODEL.ROI_MASK_HEAD.POOLER_SAMPLING_RATIO
        pooler = Pooler(
            output_size=(resolution, resolution),
            scales=scales,
            sampling_ratio=sampling_ratio,
        )
        input_size = cfg.MODEL.BACKBONE.OUT_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

        next_feature = input_size
        self.blocks = []
        for layer_idx, layer_features in enumerate(layers, 1):
            layer_name = "mask_fcn{}".format(layer_idx)
            module = make_conv3x3(next_feature, layer_features, 
                dilation=dilation, stride=1, use_gn=use_gn
            )
            self.add_module(layer_name, module)
            next_feature = layer_features
            self.blocks.append(layer_name) 
Example #11
Source File: roi_mask_feature_extractors.py    From NAS-FCOS with BSD 2-Clause "Simplified" License 5 votes vote down vote up
def __init__(self, cfg):
        """
        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__()

        resolution = cfg.MODEL.ROI_MASK_HEAD.POOLER_RESOLUTION
        scales = cfg.MODEL.ROI_MASK_HEAD.POOLER_SCALES
        sampling_ratio = cfg.MODEL.ROI_MASK_HEAD.POOLER_SAMPLING_RATIO
        pooler = Pooler(
            output_size=(resolution, resolution),
            scales=scales,
            sampling_ratio=sampling_ratio,
        )
        input_size = cfg.MODEL.BACKBONE.OUT_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

        next_feature = input_size
        self.blocks = []
        for layer_idx, layer_features in enumerate(layers, 1):
            layer_name = "mask_fcn{}".format(layer_idx)
            module = make_conv3x3(next_feature, layer_features, 
                dilation=dilation, stride=1, use_gn=use_gn
            )
            self.add_module(layer_name, module)
            next_feature = layer_features
            self.blocks.append(layer_name) 
Example #12
Source File: roi_mask_feature_extractors.py    From RRPN_pytorch with MIT License 5 votes vote down vote up
def __init__(self, cfg):
        """
        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__()

        resolution = cfg.MODEL.ROI_MASK_HEAD.POOLER_RESOLUTION
        scales = cfg.MODEL.ROI_MASK_HEAD.POOLER_SCALES
        sampling_ratio = cfg.MODEL.ROI_MASK_HEAD.POOLER_SAMPLING_RATIO
        pooler = PyramidRROIAlign(
            output_size=(resolution, resolution),
            scales=scales,
        )
        input_size = cfg.MODEL.BACKBONE.OUT_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

        self.word_margin = cfg.MODEL.ROI_REC_HEAD.BOXES_MARGIN
        self.det_margin = cfg.MODEL.RRPN.GT_BOX_MARGIN

        self.rescale = self.word_margin / self.det_margin

        next_feature = input_size
        self.blocks = []
        for layer_idx, layer_features in enumerate(layers, 1):
            layer_name = "mask_fcn{}".format(layer_idx)
            module = make_conv3x3(next_feature, layer_features, 
                dilation=dilation, stride=1, use_gn=use_gn
            )
            self.add_module(layer_name, module)
            next_feature = layer_features
            self.blocks.append(layer_name) 
Example #13
Source File: roi_rec_feature_extractors.py    From RRPN_pytorch with MIT License 5 votes vote down vote up
def __init__(self, cfg):
        """
        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__()

        resolution = cfg.MODEL.ROI_MASK_HEAD.POOLER_RESOLUTION
        scales = cfg.MODEL.ROI_MASK_HEAD.POOLER_SCALES
        sampling_ratio = cfg.MODEL.ROI_MASK_HEAD.POOLER_SAMPLING_RATIO
        pooler = Pooler(
            output_size=(resolution, resolution),
            scales=scales,
            sampling_ratio=sampling_ratio
        )
        input_size = cfg.MODEL.BACKBONE.OUT_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

        next_feature = input_size
        self.blocks = []
        for layer_idx, layer_features in enumerate(layers, 1):
            layer_name = "mask_fcn{}".format(layer_idx)
            module = make_conv3x3(next_feature, layer_features, 
                dilation=dilation, stride=1, use_gn=use_gn
            )
            self.add_module(layer_name, module)
            next_feature = layer_features
            self.blocks.append(layer_name) 
Example #14
Source File: roi_mask_feature_extractors.py    From RRPN_pytorch with MIT License 5 votes vote down vote up
def __init__(self, cfg):
        """
        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__()

        resolution = cfg.MODEL.ROI_MASK_HEAD.POOLER_RESOLUTION
        scales = cfg.MODEL.ROI_MASK_HEAD.POOLER_SCALES
        sampling_ratio = cfg.MODEL.ROI_MASK_HEAD.POOLER_SAMPLING_RATIO
        pooler = Pooler(
            output_size=(resolution, resolution),
            scales=scales,
            sampling_ratio=sampling_ratio,
        )
        input_size = cfg.MODEL.BACKBONE.OUT_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

        next_feature = input_size
        self.blocks = []
        for layer_idx, layer_features in enumerate(layers, 1):
            layer_name = "mask_fcn{}".format(layer_idx)
            module = make_conv3x3(next_feature, layer_features, 
                dilation=dilation, stride=1, use_gn=use_gn
            )
            self.add_module(layer_name, module)
            next_feature = layer_features
            self.blocks.append(layer_name) 
Example #15
Source File: roi_mask_feature_extractors.py    From DF-Traffic-Sign-Identification with MIT License 5 votes vote down vote up
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__()

        resolution = cfg.MODEL.ROI_MASK_HEAD.POOLER_RESOLUTION
        scales = cfg.MODEL.ROI_MASK_HEAD.POOLER_SCALES
        sampling_ratio = cfg.MODEL.ROI_MASK_HEAD.POOLER_SAMPLING_RATIO
        pooler = Pooler(
            output_size=(resolution, resolution),
            scales=scales,
            sampling_ratio=sampling_ratio,
        )
        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

        next_feature = input_size
        self.blocks = []
        for layer_idx, layer_features in enumerate(layers, 1):
            layer_name = "mask_fcn{}".format(layer_idx)
            module = make_conv3x3(
                next_feature, layer_features,
                dilation=dilation, stride=1, use_gn=use_gn
            )
            self.add_module(layer_name, module)
            next_feature = layer_features
            self.blocks.append(layer_name)
        self.out_channels = layer_features 
Example #16
Source File: roi_mask_feature_extractors.py    From TinyBenchmark with MIT License 5 votes vote down vote up
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__()

        resolution = cfg.MODEL.ROI_MASK_HEAD.POOLER_RESOLUTION
        scales = cfg.MODEL.ROI_MASK_HEAD.POOLER_SCALES
        sampling_ratio = cfg.MODEL.ROI_MASK_HEAD.POOLER_SAMPLING_RATIO
        pooler = Pooler(
            output_size=(resolution, resolution),
            scales=scales,
            sampling_ratio=sampling_ratio,
        )
        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

        next_feature = input_size
        self.blocks = []
        for layer_idx, layer_features in enumerate(layers, 1):
            layer_name = "mask_fcn{}".format(layer_idx)
            module = make_conv3x3(
                next_feature, layer_features,
                dilation=dilation, stride=1, use_gn=use_gn
            )
            self.add_module(layer_name, module)
            next_feature = layer_features
            self.blocks.append(layer_name)
        self.out_channels = layer_features