DDGRCF / DFDet

Offical Implementation for Oriented Object Detection via Contextual Dependence Mining and Penalty-Incentive Allocation
Apache License 2.0
8 stars 2 forks source link

堆叠卷积的类型 #3

Open ZhenboZhao77 opened 1 day ago

ZhenboZhao77 commented 1 day ago
    for i in range(self.stacked_convs):
        if i < self.num_dcn:
            conv_cfg = dict(type='DCNv2', deform_groups=4)
        else:
            conv_cfg = self.conv_cfg

        chn = self.in_channels if i == 0 else self.feat_channels
        self.inter_convs.append(
            FeatureEnhanceModule(
                chn,
                self.feat_channels,
                self.feat_channels,
                3,
                stride=1,
                padding=1,
                conv_cfg=conv_cfg,
                norm_cfg=self.norm_cfg,
                use_se=True,
            ))

请问作者,我在测试时间发现 self.num_dcn始终为0 也就是说 堆叠的卷积都是普通的卷积块。而在循环中你还设计有可变性卷积但是好像并没有去用,请问是出于什么原因呢?

ZhenboZhao77 commented 1 day ago

ScaleAdaptiveHead 中代码