module moplayground.learning.training

---

function plot_mo_progress

plot_mo_progress(
    num_steps: int,
    metrics: dict,
    training_data: moplayground.learning.training.MOTrainingInfo,
    save_dir: pathlib.Path,
    run: wandb.sdk.wandb_run.Run = None
)

---

function mo_wrapper

mo_wrapper(
    env: mujoco_playground._src.mjx_env.MjxEnv,
    episode_length: int = 1000,
    action_repeat: int = 1,
    randomization_fn=None
) → Wrapper

Multi-Objective Wrapper

---

function mo_train

mo_train(
    config_yaml: dict,
    output_dir: pathlib.Path,
    env: moplayground.envs.generic.mobase.MultiObjectiveBase,
    eval_env: moplayground.envs.generic.mobase.MultiObjectiveBase,
    moppo_params,
    network_params,
    policy_init_params,
    run: wandb.sdk.wandb_run.Run
)

---

class MOTrainingInfo

MOTrainingInfo(start_time: float, times: list = , iterations: list = , paretos: list = , directives: list = , labels: list = )

method MOTrainingInfo.__init__

__init__(
    start_time: float,
    times: list = <factory>,
    iterations: list = <factory>,
    paretos: list = <factory>,
    directives: list = <factory>,
    labels: list = <factory>
) → None

---

method MOTrainingInfo.save

save(save_dir, create_time=True)