HKUDS/DeepTutor · error · ValueError

No active image-generation model is configured. Set it in Se

Error message

No active image-generation model is configured. Set it in Settings > Media Generation > Image Generation.

What it means

Raised by resolve_imagegen_runtime_config when the active profile's model entry in the model catalog has no 'model' field (empty/missing model id) for the 'imagegen' category. The resolver refuses to build a generation config without a concrete model name. It is a configuration error, thrown before any provider call.

Source

Thrown at deeptutor/services/config/provider_runtime.py:1085

        base_url=api_base,
        api_version=_as_str((profile or {}).get("api_version")) or None,
        extra_headers=_to_headers((profile or {}).get("extra_headers")),
        language=_as_str((model or {}).get("language")) or None,
    )


def resolve_imagegen_runtime_config(
    catalog: dict[str, Any] | None = None,
    *,
    service: ModelCatalogService | None = None,
) -> ImagegenConfig:
    """Resolve the active text-to-image config from the model catalog."""
    catalog_service = service or get_model_catalog_service()
    loaded = _load_catalog(catalog)
    profile, model = _active_profile_and_model(loaded, catalog_service, "imagegen")
    resolved_model = _as_str((model or {}).get("model"))
    if not resolved_model:
        raise ValueError(
            "No active image-generation model is configured. "
            "Set it in Settings > Media Generation > Image Generation."
        )

    provider = _canonical_generation_provider(
        _as_str((profile or {}).get("binding")), IMAGEGEN_PROVIDERS
    )
    spec = IMAGEGEN_PROVIDERS[provider]
    api_base = _as_str((profile or {}).get("base_url")) or spec.default_api_base
    api_key = _as_str((profile or {}).get("api_key"))
    if not api_key and spec.is_local:
        api_key = "sk-no-key-required"

    return ImagegenConfig(
        model=resolved_model,
        provider_name=provider,
        adapter=spec.adapter,
        auth_style=spec.auth_style,

View on GitHub (pinned to 3e82f13042)

Solutions

Set an active image-generation model in Settings > Media Generation > Image Generation
Edit the model catalog JSON so the active imagegen profile's model object contains a non-empty 'model' string
Guard programmatic callers with a pre-check and surface a setup prompt instead of crashing

Example fix

// before
cfg = resolve_imagegen_runtime_config()  # ValueError
// after
try:
    cfg = resolve_imagegen_runtime_config()
except ValueError as e:
    print("configure image model:", e)
    raise SystemExit(2)

Defensive patterns

Strategy: validation

Validate before calling

def imagegen_configured() -> bool:
    try:
        resolve_imagegen_runtime_config()
        return True
    except ValueError:
        return False

Try / catch

try:
    cfg = resolve_imagegen_runtime_config()
except ValueError as e:
    if "image-generation model" in str(e):
        prompt_user_to_configure("imagegen")
    raise

Prevention

Gate image-generation UI/actions behind a config pre-check
Validate the catalog after migrations: active imagegen profile must have a non-empty model string

When it happens

Trigger: Calling generate_image or the provider test hook (_test_imagegen) when Settings > Media Generation > Image Generation has no active model selected, or the catalog JSON has an imagegen profile whose model object lacks a 'model' key.

Common situations: Fresh install where media generation was never configured; catalog reset/migration wiped the model field; UI allows saving an empty model selection; tests asserting the error path (test_resolve_imagegen_config_raises_without_model).

Related errors

AI-assisted analysis of HKUDS/DeepTutor@3e82f13042 (2026-08-27). Data as JSON: /api/errors/5cb5878ec6d2d53f. Report an issue: GitHub.