42588e0b073d51124b87b66d04aa41373d462952,keras_/kerascv/models/common.py,,conv2d,#Any#Any#Any#Any#Any#Any#Any#Any#Any#Any#,158

Before Change

        in_group_channels = in_channels // groups
        out_group_channels = out_channels // groups
        group_list = []
        is_channels_first = (K.image_data_format() == "channels_first")
        for gi in range(groups):
            xi = nn.Lambda(
                (lambda z: z[:, gi * in_group_channels:(gi + 1) * in_group_channels, :, :])
                if is_channels_first else
                (lambda z: z[:, :, :, gi * in_group_channels:(gi + 1) * in_group_channels]))(x)
            xi = nn.Conv2D(
                filters=out_group_channels,
                kernel_size=kernel_size,
                strides=strides,
                padding=ke_padding,
                dilation_rate=dilation,
                use_bias=use_bias,
                name=name + "/convgroup{}".format(gi + 1))(xi)
            group_list.append(xi)
        channel_axis = 1 if is_channels_first else -1
        x = nn.concatenate(group_list, axis=channel_axis, name=name + "/concat")
        if none_batch and (x._keras_shape[0] is not None):
            x._keras_shape = (None, ) + x._keras_shape[1:]

After Change

                use_bias=use_bias,
                name=name + "/convgroup{}".format(gi + 1))(xi)
            group_list.append(xi)
        x = nn.concatenate(group_list, axis=get_channel_axis(), name=name + "/concat")
        if none_batch and (x._keras_shape[0] is not None):
            x._keras_shape = (None, ) + x._keras_shape[1:]

Instances