opendatalab/MinerU · error · ValueError

Unsupported lmdeploy backend: {lm_backend}

Error message

Unsupported lmdeploy backend: {lm_backend}

What it means

Raised when the lmdeploy engine backend name supplied via the `lmdeploy_backend` kwarg is neither 'pytorch' nor 'turbomind'. This validation runs only when MINERU_LMDEPLOY_BACKEND is unset; the value decides whether a PytorchEngineConfig or TurbomindEngineConfig is built.

Source

Thrown at mineru/backend/vlm/vlm_analyze.py:197

                        except ImportError:
                            raise ImportError("Please install lmdeploy to use the lmdeploy-engine backend.")
                        if "cache_max_entry_count" not in kwargs:
                            kwargs["cache_max_entry_count"] = 0.5

                        device_type = os.getenv("MINERU_LMDEPLOY_DEVICE", "")
                        if device_type == "":
                            if "lmdeploy_device" in kwargs:
                                device_type = kwargs.pop("lmdeploy_device")
                                if device_type not in ["cuda", "ascend", "maca", "camb"]:
                                    raise ValueError(f"Unsupported lmdeploy device type: {device_type}")
                            else:
                                device_type = "cuda"
                        lm_backend = os.getenv("MINERU_LMDEPLOY_BACKEND", "")
                        if lm_backend == "":
                            if "lmdeploy_backend" in kwargs:
                                lm_backend = kwargs.pop("lmdeploy_backend")
                                if lm_backend not in ["pytorch", "turbomind"]:
                                    raise ValueError(f"Unsupported lmdeploy backend: {lm_backend}")
                            else:
                                lm_backend = set_lmdeploy_backend(device_type)
                        logger.info(f"lmdeploy device is: {device_type}, lmdeploy backend is: {lm_backend}")

                        if lm_backend == "pytorch":
                            kwargs["device_type"] = device_type
                            backend_config = PytorchEngineConfig(**kwargs)
                        elif lm_backend == "turbomind":
                            backend_config = TurbomindEngineConfig(**kwargs)
                        else:
                            raise ValueError(f"Unsupported lmdeploy backend: {lm_backend}")

                        log_level = 'ERROR'
                        from lmdeploy.utils import get_logger
                        lm_logger = get_logger('lmdeploy')
                        lm_logger.setLevel(log_level)
                        if os.getenv('TM_LOG_LEVEL') is None:
                            os.environ['TM_LOG_LEVEL'] = log_level

View on GitHub (pinned to 4fe4bde114)

Solutions

1. Use exactly 'pytorch' or 'turbomind' for lmdeploy_backend
2. Or omit the kwarg entirely and let set_lmdeploy_backend(device_type) pick the right engine for your device
3. Or set the env var MINERU_LMDEPLOY_BACKEND=pytorch (or turbomind) instead
4. Check for trailing whitespace/casing: pass lmdeploy_backend.strip().lower()

Example fix

# before
analyzer = MineVlmAnalyzer(backend='vlm-engine', lmdeploy_backend='PytorchEngine')

# after
analyzer = MineVlmAnalyzer(backend='vlm-engine')  # auto-select via set_lmdeploy_backend(device_type)

SUPPORTED_LMDEPLOY_BACKENDS = {"pytorch", "turbomind"}

if lmdeploy_backend is not None:
    assert lmdeploy_backend.strip().lower() in SUPPORTED_LMDEPLOY_BACKENDS, (
        f"lmdeploy_backend must be one of {SUPPORTED_LMDEPLOY_BACKENDS}"
    )
# best: omit lmdeploy_backend and let mineru auto-select per device

try:
    analyzer = MineVlmAnalyzer(backend="vlm-engine", lmdeploy_backend=lm_backend)
except ValueError as e:
    if "Unsupported lmdeploy backend" in str(e):
        lm_backend = None  # fall back to auto-selection via set_lmdeploy_backend
        analyzer = MineVlmAnalyzer(backend="vlm-engine")
    else:
        raise

When it happens

Trigger: Passing lmdeploy_backend='pytorch-engine', 'Pytorch', 'triton', or any string other than pytorch/turbomind in the analyzer kwargs while MINERU_LMDEPLOY_BACKEND is empty.

Common situations: Version drift: configs written for older mineru releases that accepted different backend spellings; users assuming the kwarg accepts the same names as the lmdeploy CLI; copy-paste from lmdeploy docs where backend names differ.

Related errors

• Unsupported lmdeploy device type: {device_type}
• Unsupported lmdeploy backend: {lm_backend}
• backend={backend} requires server_url
• CUDA is not available.
• Unsupported operating system.

AI-assisted analysis of opendatalab/MinerU@4fe4bde114 (2026-08-14). Data as JSON: /api/errors/eda1ec2ca728a5bf. Report an issue: GitHub.