keras-team/keras · error · ValueError

Invalid mode '{mode}'. Supported modes are: 'full', 'activat

Error message

Invalid mode '{mode}'. Supported modes are: 'full', 'activations', 'larger_than', 'list_of_layers', or  None.

What it means

Error "Invalid mode '{mode}'. Supported modes are: 'full', 'activations', 'larger_than', 'list_of_layers', or None." thrown in keras-team/keras.

Source

Thrown at keras/src/backend/common/remat.py:88

        output1 = layer1(input_tensor)  # layer1 is fully rematerialized
        with RematScope(mode="larger_than", output_size_threshold=512):
            layer2 = keras.layers.Conv2D(32, (3, 3))
            output2 = layer2(output1) # layer2 is conditionally rematerialized
            # if output > 512
    ```
    """

    def __init__(
        self, mode="full", output_size_threshold=1024, layer_names=None
    ):
        if mode not in {
            "full",
            "activations",
            "larger_than",
            "list_of_layers",
            None,
        }:
            raise ValueError(
                f"Invalid mode '{mode}'. Supported modes are: "
                "'full', 'activations', 'larger_than', 'list_of_layers', or "
                " None."
            )
        self.mode = mode
        self.output_size_threshold = output_size_threshold
        self.layer_names = layer_names or []
        self._pop_on_exit = False

    def __enter__(self):
        remat_scope_stack = global_state.get_global_attribute(
            "remat_scope_stack", default=[], set_to_default=True
        )
        remat_scope_stack.append(self)
        self._pop_on_exit = True
        return self

    def __exit__(self, *args, **kwargs):

View on GitHub (pinned to 7a34a03db6)

When it happens

Trigger: Thrown at keras/src/backend/common/remat.py:88 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/9a7162a0abf209ad. Report an issue: GitHub.