Multilingual Support

[uraurora]

version: gdm-concordia = "^2.1.0"

I hope the agent can support multiple languages. I don't know how to set the global model language or the language of individual agents, and I'm wondering if there are methods other than modifying the prompt.

[jzleibo]

modifying prompts is definitely the way to go

[uraurora]

modifying prompts is definitely the way to go

OK, I see. So which parameter is the best practice for modifying the agent? I noticed there are some custom keys in the entity's params. Should the prompt words for setting the agent's output language also be handled this way? Or if I want to modify other properties:

instances = [
    prefab_lib.InstanceConfig(
        prefab='basic__Entity',
        role=prefab_lib.Role.ENTITY,
        params={
            'name': 'Oliver Cromwell',
            'goal': 'become lord protector',
        },
    ),
    prefab_lib.InstanceConfig(
        prefab='basic__Entity',
        role=prefab_lib.Role.ENTITY,
        params={
            'name': 'King Charles I',
            'goal': 'avoid execution for treason',
        },
    ),
    prefab_lib.InstanceConfig(
        prefab='generic__GameMaster',
        role=prefab_lib.Role.GAME_MASTER,
        params={
            'name': 'default rules',
            # Comma-separated list of thought chain steps.
            'extra_event_resolution_steps': '',
        },
    ),
    prefab_lib.InstanceConfig(
        prefab='formative_memories_initializer__GameMaster',
        role=prefab_lib.Role.INITIALIZER,
        params={
            'name': 'initial setup rules',
            'next_game_master_name': 'default rules',
            'shared_memories': [
                'The king was captured by Parliamentary forces in 1646.',
                'Charles I was tried for treason and found guilty.',
            ],
        },
    ),
]
#
config = prefab_lib.Config(
    default_premise='Today is January 29, 1649.',
    default_max_steps=5,
    prefabs=prefabs,
    instances=instances,
)

[apexpdl]

Hi! I tried implementing this and added a short how-to guide for multilingual support, showing how to enforce per-agent output languages via prompt customization. I’ve opened a PR (#182) for review. Happy to adjust anything if it helps. Thanks!