hiyouga/LlamaFactory · error · ValueError

kernel_config.name must contain at least one kernel name.

Error message

kernel_config.name must contain at least one kernel name.

What it means

After splitting kernel_config['name'] on commas, at least one non-empty kernel name must remain. A string of only commas/whitespace (e.g. "", " , ") raises this ValueError, distinguishing 'configured but empty' from 'mis-typed'.

Source

Thrown at src/llamafactory/v1/plugins/model_plugins/kernels/interface.py:52

def _apply_auto_kernels(model: HFModel, **kwargs) -> HFModel:
    device_type = get_current_accelerator().type
    for kernel_name in _AUTO_KERNELS.get(device_type, ()):
        model = KernelPlugin(kernel_name).apply(model=model, **kwargs)

    return model


def apply_kernels(model: HFModel, config: dict[str, Any], require_logits: bool = False) -> HFModel:
    """Apply the comma-separated kernel names selected by ``kernel_config.name``."""
    kernel_names = config.get("name")
    if not isinstance(kernel_names, str):
        raise TypeError("kernel_config.name must be a string.")

    names = [name.strip() for name in kernel_names.split(",") if name.strip()]
    if not names:
        raise ValueError("kernel_config.name must contain at least one kernel name.")

    for name in names:
        if name == "auto":
            model = _apply_auto_kernels(model=model, config=config, require_logits=require_logits)
        else:
            model = KernelPlugin(name).apply(model=model, config=config, require_logits=require_logits)

    return model


def apply_v1_kernels(model: HFModel, use_v1_kernels: bool) -> HFModel:
    """Apply v1 automatic kernels for the transitional v0 ``use_v1_kernels`` option."""
    if not use_v1_kernels:
        return model

    return apply_kernels(model, {"name": "auto"})

View on GitHub (pinned to f28afaf635)

Solutions

1. Provide at least one valid kernel name or the special "auto".
2. If kernels are optional in your flow, skip calling apply_kernels entirely when the name would be empty.
3. Add a config lint step that rejects empty kernel_config.name.

Example fix

# before
kernel_config: {"name": ""}

# after
kernel_config: {"name": "auto"}  # or omit the call

name = kernel_config.get('name', '')
names = [n.strip() for n in name.split(',') if n.strip()]
assert names, 'kernel_config.name resolved to zero kernels'

def resolves_to_kernels(cfg: dict) -> bool:
    """True when cfg['name'] yields at least one non-empty kernel name."""
    return bool([n for n in cfg.get('name', '').split(',') if n.strip()])

try:
    apply_kernels(model, cfg)
except ValueError as e:
    if 'at least one kernel name' in str(e):
        cfg['name'] = 'auto'
        apply_kernels(model, cfg)
    else:
        raise

When it happens

Trigger: kernel_config name set to "" or " , " via YAML interpolation, env var expansion, or templating that left the value empty.

Common situations: Templated configs (Jinja/envsubst) where the kernel list variable is unset; YAML `name: "${KERNELS:-}" producing empty string; trailing-comma edits.

Related errors

• kernel_config.name must be a string.
• Unknown Liger op(s) {sorted(ops)} for model_type={model_type
• Unknown mixing strategy: {data_args.mix_strategy}.
• Cannot specify `val_size` if `eval_dataset` is not None.
• `kt_model_max_length` must be a positive integer.

AI-assisted analysis of hiyouga/LlamaFactory@f28afaf635 (2026-08-14). Data as JSON: /api/errors/0c502de588c41352. Report an issue: GitHub.