How to use the csbdeep.utils.move_channel_for_backend function in csbdeep

def to_tensor(x,channel=None,single_sample=True):
    if single_sample:
        x = x[np.newaxis]
        if channel is not None and channel >= 0:
            channel += 1
    if channel is None:
        x, channel = np.expand_dims(x,-1), -1
    return move_channel_for_backend(x,channel)

assert X.shape[0] == Y.shape[0]
    assert 0 < n_images <= X.shape[0]
    assert 0 <= validation_split < 1

    X, Y = X[:n_images], Y[:n_images]
    channel = axes_dict(axes)['C']

    if validation_split > 0:
        n_val   = int(round(n_images * validation_split))
        n_train = n_images - n_val
        assert 0 < n_val and 0 < n_train
        X_t, Y_t = X[-n_val:],  Y[-n_val:]
        X,   Y   = X[:n_train], Y[:n_train]
        assert X.shape[0] == n_train and X_t.shape[0] == n_val
        X_t = move_channel_for_backend(X_t,channel=channel)
        Y_t = move_channel_for_backend(Y_t,channel=channel)

    X = move_channel_for_backend(X,channel=channel)
    Y = move_channel_for_backend(Y,channel=channel)

    axes = axes.replace('C','') # remove channel
    if backend_channels_last():
        axes = axes+'C'
    else:
        axes = axes[:1]+'C'+axes[1:]

    data_val = (X_t,Y_t) if validation_split > 0 else None

    if verbose:
        ax = axes_dict(axes)
        n_train, n_val = len(X), len(X_t) if validation_split>0 else 0
        image_size = tuple( X.shape[ax[a]] for a in axes if a in 'TZYX' )

assert 0 <= validation_split < 1

    X, Y = X[:n_images], Y[:n_images]
    channel = axes_dict(axes)['C']

    if validation_split > 0:
        n_val   = int(round(n_images * validation_split))
        n_train = n_images - n_val
        assert 0 < n_val and 0 < n_train
        X_t, Y_t = X[-n_val:],  Y[-n_val:]
        X,   Y   = X[:n_train], Y[:n_train]
        assert X.shape[0] == n_train and X_t.shape[0] == n_val
        X_t = move_channel_for_backend(X_t,channel=channel)
        Y_t = move_channel_for_backend(Y_t,channel=channel)

    X = move_channel_for_backend(X,channel=channel)
    Y = move_channel_for_backend(Y,channel=channel)

    axes = axes.replace('C','') # remove channel
    if backend_channels_last():
        axes = axes+'C'
    else:
        axes = axes[:1]+'C'+axes[1:]

    data_val = (X_t,Y_t) if validation_split > 0 else None

    if verbose:
        ax = axes_dict(axes)
        n_train, n_val = len(X), len(X_t) if validation_split>0 else 0
        image_size = tuple( X.shape[ax[a]] for a in axes if a in 'TZYX' )
        n_dim = len(image_size)
        n_channel_in, n_channel_out = X.shape[ax['C']], Y.shape[ax['C']]

X, Y = X[:n_images], Y[:n_images]
    channel = axes_dict(axes)['C']

    if validation_split > 0:
        n_val   = int(round(n_images * validation_split))
        n_train = n_images - n_val
        assert 0 < n_val and 0 < n_train
        X_t, Y_t = X[-n_val:],  Y[-n_val:]
        X,   Y   = X[:n_train], Y[:n_train]
        assert X.shape[0] == n_train and X_t.shape[0] == n_val
        X_t = move_channel_for_backend(X_t,channel=channel)
        Y_t = move_channel_for_backend(Y_t,channel=channel)

    X = move_channel_for_backend(X,channel=channel)
    Y = move_channel_for_backend(Y,channel=channel)

    axes = axes.replace('C','') # remove channel
    if backend_channels_last():
        axes = axes+'C'
    else:
        axes = axes[:1]+'C'+axes[1:]

    data_val = (X_t,Y_t) if validation_split > 0 else None

    if verbose:
        ax = axes_dict(axes)
        n_train, n_val = len(X), len(X_t) if validation_split>0 else 0
        image_size = tuple( X.shape[ax[a]] for a in axes if a in 'TZYX' )
        n_dim = len(image_size)
        n_channel_in, n_channel_out = X.shape[ax['C']], Y.shape[ax['C']]

n_images = X.shape[0]
    assert X.shape[0] == Y.shape[0]
    assert 0 < n_images <= X.shape[0]
    assert 0 <= validation_split < 1

    X, Y = X[:n_images], Y[:n_images]
    channel = axes_dict(axes)['C']

    if validation_split > 0:
        n_val   = int(round(n_images * validation_split))
        n_train = n_images - n_val
        assert 0 < n_val and 0 < n_train
        X_t, Y_t = X[-n_val:],  Y[-n_val:]
        X,   Y   = X[:n_train], Y[:n_train]
        assert X.shape[0] == n_train and X_t.shape[0] == n_val
        X_t = move_channel_for_backend(X_t,channel=channel)
        Y_t = move_channel_for_backend(Y_t,channel=channel)

    X = move_channel_for_backend(X,channel=channel)
    Y = move_channel_for_backend(Y,channel=channel)

    axes = axes.replace('C','') # remove channel
    if backend_channels_last():
        axes = axes+'C'
    else:
        axes = axes[:1]+'C'+axes[1:]

    data_val = (X_t,Y_t) if validation_split > 0 else None

    if verbose:
        ax = axes_dict(axes)
        n_train, n_val = len(X), len(X_t) if validation_split>0 else 0