mudler/LocalAI · error · ValueError

resolution must be 480p or 720p

Error message

resolution must be 480p or 720p

What it means

ValueError raised during longcat-video model loading when the 'resolution' option is anything other than '480p' or '720p' (case-insensitive, default '480p'). LongCat-Video pipelines are compiled/trained at fixed resolutions, so the value is validated and normalized before being stored into self.options.

Source

Thrown at backend/python/longcat-video/backend.py:167

            if not self.torch.cuda.is_available():
                return self._fail(
                    context,
                    grpc.StatusCode.FAILED_PRECONDITION,
                    "longcat-video requires an NVIDIA CUDA GPU",
                )
            if request.TensorParallelSize > 1:
                return self._fail(
                    context,
                    grpc.StatusCode.UNIMPLEMENTED,
                    "longcat-video currently supports one GPU per backend process",
                )
            self._import_runtime()

            attention_name = str(options.get("attention_backend", "sdpa")).lower()
            attention_overrides(attention_name)
            resolution = str(options.get("resolution", "480p")).lower()
            if resolution not in {"480p", "720p"}:
                raise ValueError("resolution must be 480p or 720p")

            use_distill_default = model_kind == MODEL_KIND_AVATAR
            use_distill = require_bool(
                options.get("use_distill", use_distill_default),
                "use_distill",
            )
            use_int8 = require_bool(options.get("use_int8", False), "use_int8")
            if model_kind == MODEL_KIND_BASE and use_int8:
                raise ValueError(
                    "use_int8 is supported only by LongCat-Video-Avatar-1.5"
                )

            self.options = {
                **options,
                "attention_backend": attention_name,
                "resolution": resolution,
                "use_distill": use_distill,
                "use_int8": use_int8,

View on GitHub (pinned to 44413a9d06)

Solutions

1. Set resolution to exactly '480p' or '720p' (any casing) or omit it to get the 480p default
2. Remove custom resolution strings like widthxheight from longcat-video options

Example fix

# before
options:
  resolution: 1080p

# after
options:
  resolution: 720p

VALID_RESOLUTIONS = {"480p", "720p"}

def normalize_resolution(options: dict) -> dict:
    res = str(options.get("resolution", "480p")).strip().lower()
    if res not in VALID_RESOLUTIONS:
        raise ValueError(f"resolution must be one of {sorted(VALID_RESOLUTIONS)}, got {options.get('resolution')!r}")
    return {**options, "resolution": res}

def is_valid_resolution(value) -> bool:
    return isinstance(value, str) and value.strip().lower() in {"480p", "720p"}

try:
    stub.LoadModel(model_options)
except grpc.RpcError as e:
    if "resolution must be" in (e.details() or ""):
        model_options["options"]["resolution"] = "480p"
        stub.LoadModel(model_options)
    else:
        raise

When it happens

Trigger: LoadModel with options {"resolution": "1080p"} or "480" or "hd"; passing a numeric resolution like "854x480"; a typo such as "720P " is fine (lowercased) but "720" is not.

Common situations: User assumes arbitrary resolutions are supported because other video backends accept width/height; copy-pasting resolution strings from diffusion configs.

Related errors

• use_int8 is supported only by LongCat-Video-Avatar-1.5
• num_frames must not be negative
• base_model must point to a LongCat-Video checkpoint
• request needs {segments} avatar segments, but max_segments i
• {name} must be true or false

AI-assisted analysis of mudler/LocalAI@44413a9d06 (2026-08-15). Data as JSON: /api/errors/c6eb535d920a796a. Report an issue: GitHub.