keras-team/keras · error · ValueError

Unknown fill_mode: {fill_mode}

Error message

Unknown fill_mode: {fill_mode}

What it means

Error "Unknown fill_mode: {fill_mode}" thrown in keras-team/keras.

Source

Thrown at keras/src/backend/tensorflow/image.py:712

                coords, dtype=tf.bool
            )
        elif fill_mode in ["mirror", "reflect"]:
            coords = tf.abs(coords)
            size_2 = size * 2
            mod = tf.math.mod(coords, size_2)
            under = mod < size
            over = ~under
            # reflect mode is same as mirror for under
            coords = tf.where(under, mod, size_2 - mod)
            # for reflect mode, adjust the over case
            if fill_mode == "reflect":
                coords = tf.where(over, coords - 1, coords)
            return coords, tf.ones_like(coords, dtype=tf.bool)
        elif fill_mode == "wrap":
            coords = tf.math.mod(coords, size)
            return coords, tf.ones_like(coords, dtype=tf.bool)
        else:
            raise ValueError(f"Unknown fill_mode: {fill_mode}")

    valid_1d_interpolations = []
    for coordinate, size in zip(coordinate_arrs, input_arr.shape):
        interp_nodes = interp_fun(coordinate)
        valid_interp = []
        for index, weight in interp_nodes:
            safe_index, valid = process_coordinates(index, size)
            valid_interp.append((safe_index, valid, weight))
        valid_1d_interpolations.append(valid_interp)

    outputs = []
    for items in itertools.product(*valid_1d_interpolations):
        indices, validities, weights = zip(*items)
        indices = tf.transpose(tf.stack(indices))

        gathered = tf.transpose(tf.gather_nd(input_arr, indices))

        # Cast to computation dtype early to avoid type issues

View on GitHub (pinned to 7a34a03db6)

When it happens

Trigger: Thrown at keras/src/backend/tensorflow/image.py:712 when the library encounters an invalid state.

Common situations: See trigger scenarios.

AI-assisted analysis of keras-team/keras@7a34a03db6 (2026-08-25). Data as JSON: /api/errors/6b3c6bd4297bc919. Report an issue: GitHub.