# tff.learning.programs.train_model_with_vizier

<br />

|---|
| [![](https://www.tensorflow.org/images/GitHub-Mark-32px.png) View source on GitHub](https://github.com/tensorflow/federated/blob/v0.87.0 Version 2.0, January 2004 Licensed under the Apache License, Version 2.0 (the) |

Trains and tunes a federated model using Vizier.

    tff.learning.programs.train_model_with_vizier(
        *,
        study,
        total_trials,
        num_parallel_trials=1,
        update_hparams,
        train_model_program_logic,
        train_process_factory,
        train_data_source,
        total_rounds,
        num_clients,
        program_state_manager_factory,
        model_output_manager_factory,
        train_metrics_manager_factory=None,
        evaluation_manager_factory,
        evaluation_periodicity
    )

<br />

<br />

<br />

| ## Args ||
|---|---|
| `study` | The Vizier study to use to to tune `train_model_program_logic`. |
| `total_trials` | The number of Vizier trials. |
| `num_parallel_trials` | The number of Vizier trials to be evaluated in parallel. Default is 1. |
| `update_hparams` | A [`tff.Computation`](https://www.tensorflow.org/federated/api_docs/python/tff/Computation) to use to update the models hparams using a trials parameters. |
| `train_model_program_logic` | The program logic to use for training and evaluating the model. |
| `train_process_factory` | A factory for creating [`tff.learning.templates.LearningProcess`](https://www.tensorflow.org/federated/api_docs/python/tff/learning/templates/LearningProcess) to run for training. |
| `train_data_source` | A [`tff.program.FederatedDataSource`](https://www.tensorflow.org/federated/api_docs/python/tff/program/FederatedDataSource) which returns client data used during training. |
| `total_rounds` | The number of rounds of training. |
| `num_clients` | The number of clients per round of training. |
| `program_state_manager_factory` | A factory for creating [`tff.program.ProgramStateManager`](https://www.tensorflow.org/federated/api_docs/python/tff/program/ProgramStateManager)s for each trail. |
| `model_output_manager_factory` | A factory for creating [`tff.program.ReleaseManager`](https://www.tensorflow.org/federated/api_docs/python/tff/program/ReleaseManager)s used to release the model. |
| `train_metrics_manager_factory` | A factory for creating [`tff.program.ReleaseManager`](https://www.tensorflow.org/federated/api_docs/python/tff/program/ReleaseManager)s used to release training metrics for each trail. |
| `evaluation_manager_factory` | A factory for creating [`tff.learning.programs.EvaluationManager`](https://www.tensorflow.org/federated/api_docs/python/tff/learning/programs/EvaluationManager)s for each trail. |
| `evaluation_periodicity` | Either a integer number of rounds or `datetime.timedelta` to await before sending a new training checkpoint to `evaluation_manager.start_evaluation`. |

<br />