keras-team/keras · error · ValueError

For NASNet-A models, the `penultimate_filters` must be a…

Error message

For NASNet-A models, the `penultimate_filters` must be a multiple of 24 * (`filter_multiplier` ** 2). Current value: {penultimate_filters}

What it means

NASNet family constraint in keras/src/applications/nasnet.py: the architecture's penultimate filter count must satisfy the divisibility relation penultimate_filters % (24 * filter_multiplier**2) == 0 for the reduction splits to tile evenly. A custom penultimate_filters value that violates this (e.g. for NASNetMobile: multiple of 24*4=96) makes block construction impossible, so the builder rejects it.

Solutions

1. Use penultimate_filters=4096 with filter_multiplier=24 (NASNetLarge) or 1056 with 8 (NASNetMobile)
2. Choose a penultimate_filters value divisible by 24 * filter_multiplier**2
3. Keep default architecture parameters instead of customizing filter counts

When it happens

Trigger: Thrown at keras/src/applications/nasnet.py:181 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/02f26b579b76483e. Report an issue: GitHub.

Appendix: source

Thrown at keras/src/applications/nasnet.py:181

            "However your settings specify the default "
            'data format "channels_first" (channels, width, height).'
            ' You should set `image_data_format="channels_last"` '
            "in your Keras config located at ~/.keras/keras.json. "
            "The model being returned right now will expect inputs "
            'to follow the "channels_last" data format.',
            stacklevel=2,
        )

    if input_tensor is None:
        img_input = layers.Input(shape=input_shape)
    else:
        if not backend.is_keras_tensor(input_tensor):
            img_input = layers.Input(tensor=input_tensor, shape=input_shape)
        else:
            img_input = input_tensor

    if penultimate_filters % (24 * (filter_multiplier**2)) != 0:
        raise ValueError(
            "For NASNet-A models, the `penultimate_filters` must be a multiple "
            "of 24 * (`filter_multiplier` ** 2). "
            f"Current value: {penultimate_filters}"
        )

    channel_dim = 1 if backend.image_data_format() == "channels_first" else -1
    filters = penultimate_filters // 24

    x = layers.Conv2D(
        stem_block_filters,
        (3, 3),
        strides=(2, 2),
        padding="valid",
        use_bias=False,
        name="stem_conv1",
        kernel_initializer="he_normal",
    )(img_input)

View on GitHub (pinned to 7a34a03db6)