mlflow/mlflow · error · TypeError

Argument 'mlflow_model' should be mlflow.models.Model, got '

Error message

Argument 'mlflow_model' should be mlflow.models.Model, got '{type(mlflow_model)}'

What it means

FileStore.record_logged_model requires mlflow_model to be an instance of mlflow.models.Model and raises a plain TypeError for anything else. It is an internal API type check to prevent writing invalid logged-model tags.

Source

Thrown at mlflow/store/tracking/file_store.py:1281

                        model_id=metric.model_id,
                        run_id=run_id,
                        metric=metric,
                    )
            for tag in tags:
                # NB: If the tag run name value is set, update the run info to assure
                # synchronization.
                if tag.key == MLFLOW_RUN_NAME:
                    run_status = RunStatus.from_string(run_info.status)
                    self.update_run_info(run_id, run_status, run_info.end_time, tag.value)
                self._set_run_tag(run_info, tag)
        except Exception as e:
            raise MlflowException(e, INTERNAL_ERROR)

    def record_logged_model(self, run_id, mlflow_model):
        from mlflow.models import Model

        if not isinstance(mlflow_model, Model):
            raise TypeError(
                f"Argument 'mlflow_model' should be mlflow.models.Model, got '{type(mlflow_model)}'"
            )
        _validate_run_id(run_id)
        run_info = self._get_run_info(run_id)
        check_run_is_active(run_info)
        model_dict = mlflow_model.get_tags_dict()
        run_info = self._get_run_info(run_id)
        path = self._get_tag_path(run_info.experiment_id, run_info.run_id, MLFLOW_LOGGED_MODELS)
        if os.path.exists(path):
            with open(path) as f:
                model_list = json.loads(f.read())
        else:
            model_list = []
        tag = RunTag(MLFLOW_LOGGED_MODELS, json.dumps(model_list + [model_dict]))

        try:
            self._set_run_tag(run_info, tag)
        except Exception as e:

View on GitHub (pinned to 6a27f2decc)

Solutions

1. Pass an actual mlflow.models.Model instance (e.g. the object from Model.load or the model you saved)
2. If you have a dict, construct Model.from_dict(model_dict) before calling
3. Check you are importing Model from mlflow.models, not a similarly named registry entity class
4. Use the public mlflow.log_model / MlflowClient.log_logged_model APIs instead of the internal record_logged_model

Example fix

// before
from mlflow.entities import Model
store.record_logged_model(run_id, model_dict)
// after
from mlflow.models import Model
store.record_logged_model(run_id, Model.from_dict(model_dict))

from mlflow.models import Model
if not isinstance(mlflow_model, Model):
    raise TypeError("record_logged_model requires mlflow.models.Model")

from mlflow.models import Model
def is_mlflow_model(obj) -> bool:
    return isinstance(obj, Model)

try:
    store.record_logged_model(run_id, model)
except TypeError as e:
    model = Model.from_dict(model_as_dict)
    store.record_logged_model(run_id, model)

When it happens

Trigger: Calling FileStore.record_logged_model (or a code path that forwards into it) with a dict, an mlflow.entities.model_registry.Model, a LoggedModel, or any object that is not mlflow.models.Model.

Common situations: Hand-rolled logging code passing a model dict or registry model instead of a loaded mlflow.models.Model; mixing up mlflow.models.Model with mlflow.entities.Model in custom integrations; version changes where internal helpers expect Model.

Understand the failure class

Background: "Wrong argument type", "must be a string", "expected Array or Prism::Scope": TypeError and ArgumentError when a library receives a value of the wrong type — this error's family across 28 libraries.

Related errors

• The `requestPreview` parameter must be a string.
• The `responsePreview` parameter must be a string.
• trackingUri must be a string
• experimentId must be a string
• Logged model {model_id} is not in `deleted` lifecycle stage.

AI-assisted analysis of mlflow/mlflow@6a27f2decc (2026-08-29). Data as JSON: /api/errors/ad29b160b08bdd47. Report an issue: GitHub.