keras-team/keras · error · ValueError

adapt() received an empty iterable (no batches). Expected at

Error message

adapt() received an empty iterable (no batches). Expected at least one batch. Pass a non-empty iterable of arrays or tensors, e.g. layer.adapt([x]) or layer.adapt(list_of_batches).

What it means

Normalization.adapt() iterates over batches; when handed a generic iterable it pulls the first batch to infer the input shape. If the iterator is immediately exhausted (first_batch is None), it raises this ValueError instead of silently producing garbage statistics.

Source

Thrown at keras/src/layers/preprocessing/normalization.py:301

                    if isinstance(element_spec, tuple)
                    else element_spec
                )
                return tuple(x_spec.shape)

            input_shape = get_input_shape(data)
            if len(input_shape) == 1:
                data = data.batch(128)
                input_shape = get_input_shape(data)
        elif isinstance(data, PyDataset):
            input_shape = _extract_batch(data[0]).shape
        elif hasattr(data, "__iter__"):
            data_is_iterable = True
            # Consume first batch to infer input_shape; then chain it back for
            # accumulation so we iterate over (first_batch, *rest).
            data_iter = iter(data)
            first_batch = next(data_iter, None)
            if first_batch is None:
                raise ValueError(
                    "adapt() received an empty iterable (no batches). "
                    "Expected at least one batch. Pass a non-empty iterable "
                    "of arrays or tensors, e.g. layer.adapt([x]) or "
                    "layer.adapt(list_of_batches)."
                )
            first_batch = _extract_batch(first_batch)
            input_shape = getattr(first_batch, "shape", None)
            if input_shape is None:
                raise TypeError(
                    "adapt() expects an iterable that yields arrays or "
                    "tensors with a `.shape` attribute (e.g. numpy arrays or "
                    "backend tensors). Got an element of type "
                    f"{type(first_batch).__name__}. Ensure each yielded "
                    "element is array-like with a `.shape` attribute."
                )
            input_shape = tuple(input_shape)
            data = itertools.chain([first_batch], data_iter)
        else:

View on GitHub (pinned to 7a34a03db6)

Solutions

1. Pass a non-empty iterable, e.g. layer.adapt([x]) for a single array
2. Check the data source is non-empty before adapt
3. Fix the upstream filter that emptied the dataset

Example fix

// before
layer.adapt(filtered_df[col])  # filtered empty
// after
assert len(filtered_df) > 0
layer.adapt(filtered_df[col].to_numpy())

first = next(iter(data), None)
if first is None:
    raise ValueError('cannot adapt on empty data; pass e.g. [x]')

When it happens

Trigger: layer.adapt([]), adapt(generator_that_yields_nothing), adapt(filter(...) with no matching rows), or a tf.data.Dataset of size 0.

Common situations: Empty train split after a bad filter/mask; placeholder lists during scaffolding; generators guarded by a condition that never fires.

Related errors

• adapt() expects an iterable that yields arrays or tensors wi
• Unsupported data type: {type(data)}. `adapt` supports `np.nd
• The layer was built with input_shape={self._build_input_shap
• You need to call `.adapt(dataset)` on the FeatureSpace befor
• Cannot adapt layer '{self.name}' after setting a static voca

AI-assisted analysis of keras-team/keras@7a34a03db6 (2026-08-25). Data as JSON: /api/errors/2087e8e1ece830c8. Report an issue: GitHub.