keras-team/keras · error · ValueError

Cannot concatenate features because feature '{name}' has not

Error message

Cannot concatenate features because feature '{name}' has not been encoded (it has dtype {dtype}). Consider using `output_mode='dict'`.

What it means

In output_mode='concat', FeatureSpace concatenates all features into one float tensor. If a feature still has an integer or string dtype it was never encoded (e.g. a crossing with output_mode='int' left raw), and concatenation is impossible. Set an encoding output mode or use 'dict' output.

Source

Thrown at keras/src/layers/preprocessing/feature_space.py:721

                else:
                    raise ValueError(
                        f"Feature '{name}' has `output_mode='one_hot'`. "
                        "However it isn't a standard feature and the "
                        "dimensionality of its output space is not known, "
                        "thus it cannot be one-hot encoded. "
                        "Try using `output_mode='int'`."
                    )
                if cardinality is not None:
                    encoder = layers.CategoryEncoding(
                        num_tokens=cardinality, output_mode="multi_hot"
                    )
                    self.one_hot_encoders[name] = encoder
                    feature = encoder(feature)

            if self.output_mode == "concat":
                dtype = feature.dtype
                if dtype.startswith("int") or dtype == "string":
                    raise ValueError(
                        f"Cannot concatenate features because feature '{name}' "
                        f"has not been encoded (it has dtype {dtype}). "
                        "Consider using `output_mode='dict'`."
                    )
                features_to_concat.append(feature)
            else:
                output_dict[name] = feature

        if self.output_mode == "concat":
            self.concat = TFDConcat(axis=-1)
            return self.concat(features_to_concat)
        else:
            return output_dict

    def _check_if_adapted(self):
        if not self._is_adapted:
            if not self._list_adaptable_preprocessors():
                self._is_adapted = True

View on GitHub (pinned to 7a34a03db6)

Solutions

1. Set crossing_output_mode='one_hot' (or FeatureSpace output_mode='one_hot') so cross features get encoded before concat
2. Use output_mode='dict' if you want raw per-feature outputs
3. Verify each feature's spec produces encoded (float) output when concatenating

Example fix

// before
fs = FeatureSpace(..., output_mode="concat")  # crossing_output_mode='int' default
// after
fs = FeatureSpace(..., output_mode="concat", crossing_output_mode="one_hot")
# or use output_mode="dict" for raw outputs

p = fs.preprocessors.get(name) or fs.crossers.get(name)
out = p(sample_batch)
assert not (str(out.dtype).startswith("int") or str(out.dtype) == "string"), f"{name} unencoded for concat"

def is_encoded(name, feature_space):
    return not (name in feature_space.crossers and getattr(feature_space, "crossing_output_mode", "int") == "int" and feature_space.output_mode == "concat")

catch ValueError from get_encoded_features()/__call__, then either set an encoding output_mode or switch FeatureSpace to output_mode='dict'

When it happens

Trigger: FeatureSpace(output_mode='concat') while a feature or crossing produces raw int/string output, typically crossing_output_mode='int' left at default with concat output.

Common situations: Leaving crossing_output_mode='int' while FeatureSpace output_mode='concat'; forgetting to set output_mode on the FeatureSpace so unencoded integer lookups flow into concat.

Related errors

• `adapt()` can only be called on a tf.data.Dataset or a dict
• Feature '{name}' has `output_mode='one_hot'`. Thus its prepr
• Feature '{name}' has `output_mode='one_hot'`. However it isn
• You need to call `.adapt(dataset)` on the FeatureSpace befor
• A FeatureSpace can only be called with a dict. Received: dat

AI-assisted analysis of keras-team/keras@7a34a03db6 (2026-08-25). Data as JSON: /api/errors/724a1e78ca564f99. Report an issue: GitHub.