5bfdba356f31723366a17dbc2421f345c9273199,dataset/models/tf/resnet.py,ResNet,block,#Any#Any#Any#,135

Before Change

            x_channels = cls.channels_shape(x, data_format)

            if inputs_channels != x_channels or strides > 1:
                shortcut = conv_block(inputs, "c", x_channels, 1, name="shortcut", strides=strides, **kwargs)
            else:
                shortcut = inputs

            if se_block:

After Change

        
        kwargs = cls.fill_params("body/block", **kwargs)
        filters, downsample = cls.pop(["filters", "downsample"], kwargs)
        bottleneck, bottleneck_factor = cls.pop(["bottleneck", "bottleneck_factor"], kwargs)
        resnext, resnext_factor = cls.pop(["resnext", "resnext_factor"], kwargs)
        se_block, se_factor = cls.pop(["se_block", "se_factor"], kwargs)
        activation = kwargs.get("activation")

Instances