keras-team/keras · error · ValueError

Only `vocabulary_dtype='int64'` is supported at this time. R

Error message

Only `vocabulary_dtype='int64'` is supported at this time. Received: vocabulary_dtype={vocabulary_dtype}

What it means

IntegerLookup currently supports only vocabulary_dtype='int64'. __init__ rejects any other dtype string because the underlying lookup tables are typed to int64 indices.

Source

Thrown at keras/src/layers/preprocessing/integer_lookup.py:374

                "Install it via `pip install tensorflow`."
            )
        if max_tokens is not None and max_tokens <= 1:
            raise ValueError(
                "If `max_tokens` is set for `IntegerLookup`, it must be "
                f"greater than 1. Received: max_tokens={max_tokens}"
            )
        if num_oov_indices < 0:
            raise ValueError(
                "The value of `num_oov_indices` argument for `IntegerLookup` "
                "must >= 0. Received: num_oov_indices="
                f"{num_oov_indices}"
            )
        if sparse and backend.backend() != "tensorflow":
            raise ValueError(
                "`sparse=True` can only be used with the TensorFlow backend."
            )
        if vocabulary_dtype != "int64":
            raise ValueError(
                "Only `vocabulary_dtype='int64'` is supported "
                "at this time. Received: "
                f"vocabulary_dtype={vocabulary_dtype}"
            )
        super().__init__(
            max_tokens=max_tokens,
            num_oov_indices=num_oov_indices,
            mask_token=mask_token,
            oov_token=oov_token,
            vocabulary=vocabulary,
            vocabulary_dtype=vocabulary_dtype,
            idf_weights=idf_weights,
            invert=invert,
            output_mode=output_mode,
            sparse=sparse,
            pad_to_max_tokens=pad_to_max_tokens,
            oov_method=oov_method,
            salt=salt,

View on GitHub (pinned to 7a34a03db6)

Solutions

  1. Omit vocabulary_dtype (default is int64)
  2. Cast your vocabulary array to int64: np.asarray(vocab, dtype='int64')
  3. If you need int32 output, cast after the layer: keras.ops.cast(layer(x), 'int32')

Example fix

// before
layer = IntegerLookup(vocabulary_dtype='int32')
// after
layer = IntegerLookup()  # int64 default; cast output later if needed
Defensive patterns

Strategy: validation

Validate before calling

assert vocabulary_dtype in (None, 'int64')

Prevention

When it happens

Trigger: Passing vocabulary_dtype='int32', 'float32', or anything other than 'int64' to keras.layers.IntegerLookup().

Common situations: Memory-optimization attempts to shrink vocabularies with int32; copying a config from another layer type that accepts dtype options; Keras 3 ports of older layers.

Related errors

AI-assisted analysis of keras-team/keras@7a34a03db6 (2026-08-25). Data as JSON: /api/errors/624bb00d1d74a691. Report an issue: GitHub.