Unity-Technologies/ml-agents · error · TrainerConfigError

Settings for trainer type {d_copy['trainer_type']} were not

Error message

Settings for trainer type {d_copy['trainer_type']} were not found

What it means

After determining trainer_type, the hook looks up matching hyperparameter settings in all_trainer_settings; if hyperparameters were specified but the trainer_type has no corresponding settings class registered, a KeyError is converted to TrainerConfigError.

Source

Thrown at ml-agents/mlagents/trainers/settings.py:711

        for key, val in d_copy.items():
            if attr.has(type(val)):
                # Don't convert already-converted attrs classes.
                continue
            if key == "hyperparameters":
                if "trainer_type" not in d_copy:
                    raise TrainerConfigError(
                        "Hyperparameters were specified but no trainer_type was given."
                    )
                else:
                    d_copy[key] = check_hyperparam_schedules(
                        val, d_copy["trainer_type"]
                    )
                    try:
                        d_copy[key] = strict_to_cls(
                            d_copy[key], all_trainer_settings[d_copy["trainer_type"]]
                        )
                    except KeyError:
                        raise TrainerConfigError(
                            f"Settings for trainer type {d_copy['trainer_type']} were not found"
                        )
            elif key == "max_steps":
                d_copy[key] = int(float(val))
                # In some legacy configs, max steps was specified as a float
            # elif key == "even_checkpoints":
            #     if val:
            #         d_copy["checkpoint_interval"] = int(d_copy["max_steps"] / d_copy["keep_checkpoints"])
            elif key == "trainer_type":
                if val not in all_trainer_types.keys():
                    raise TrainerConfigError(f"Invalid trainer type {val} was found")
            else:
                d_copy[key] = check_and_structure(key, val, t)
        return t(**d_copy)

    class DefaultTrainerDict(collections.defaultdict):
        def __init__(self, *args):
            # Depending on how this is called, args may have the defaultdict

View on GitHub (pinned to 3ecb446f75)

Solutions

1. Set trainer_type to one of ppo, sac, or poca (exact lowercase spelling).
2. Check for whitespace/casing typos in trainer_type.
3. Confirm the ML-Agents version supports the trainer type in your config.

Example fix

// before
trainer_type: ppo_old
// after
trainer_type: ppo

VALID = {'ppo', 'sac', 'poca'}
if cfg.get('trainer_type') not in VALID:
    raise ValueError(f'trainer_type must be one of {VALID}')

def is_known_trainer_type(v):
    return v in ('ppo', 'sac', 'poca')

from mlagents.trainers.exception import TrainerConfigError
try:
    load_config(path)
except TrainerConfigError as e:
    if 'were not found' in str(e):
        logger.error('Check trainer_type spelling: %s', e)

When it happens

Trigger: A config specifies hyperparameters plus a trainer_type whose settings are absent from all_trainer_settings — typically an invalid/unknown trainer_type string that slipped past earlier checks, or a custom/partial trainer registry.

Common situations: Typos like 'ppo ' (trailing space) or 'PPO' casing in trainer_type; version changes renaming trainer types; edited ML-Agents builds with restricted trainer registries.

Understand the failure class

Background: Invalid enum value errors: "Unknown type", "Invalid scope", "must be one of" — when a string is not on the library's allowed list — this error's family across 23 libraries.

Related errors

• Config file could not be found at {abs_path}.
• There was an error decoding Config file from {config_path}.
• Error parsing yaml file. Please check for formatting errors.
• Threshold for next lesson cannot be negative when the measur
• A non-terminal lesson does not have a completion_criteria fo

AI-assisted analysis of Unity-Technologies/ml-agents@3ecb446f75 (2026-09-02). Data as JSON: /api/errors/e96d3e53d9bc9a3a. Report an issue: GitHub.