sgl-project/sglang · error · ValueError

{name} must be a dict-like config, got {type(config)}

Error message

{name} must be a dict-like config, got {type(config)}

What it means

Raised by _normalize_config_dict during MiMoProcessor.__init__ when a config argument is not None, not a dict, and has no to_dict() method. The processor accepts dict-like configs (including HF objects exposing to_dict) and rejects anything else with the offending type in the message.

Source

Thrown at python/sglang/srt/multimodal/processors/mimo_v2.py:1486

            raise ValueError(
                f"Unrecognized image input, support local path, http url, base64 and PIL.Image, got {image}"
            )
        image = cls.to_rgb(image_obj)
        return image


class MiMoV2Processor(BaseMultimodalProcessor):
    models = [MiMoV2ForCausalLM]

    @staticmethod
    def _normalize_config_dict(config, name: str) -> dict:
        if config is None:
            return {}
        if isinstance(config, dict):
            return config
        if hasattr(config, "to_dict"):
            return config.to_dict()
        raise ValueError(f"{name} must be a dict-like config, got {type(config)}")

    @staticmethod
    def _require_config_value(config: dict, key: str):
        value = config.get(key)
        if value is None:
            raise ValueError(f"processor_config.{key} must be set for MiMo-V2")
        return value

    def _validate_placeholder_counts(
        self,
        text_parts,
        multimodal_tokens_pattern,
        image_count: int,
        video_count: int,
        audio_count: int,
    ):
        counts = {
            Modality.IMAGE: 0,

View on GitHub (pinned to 0132848349)

Solutions

1. Pass a plain dict or an HF configuration object (which has .to_dict())
2. If you have a JSON path, load it first: json.load(open(path))
3. Check {type(config)} in the message to find which argument is malformed

Example fix

# before
proc = MiMoProcessor(hf_config, server_args, _processor, processor_config='preprocessor_config.json')
# after
import json
proc = MiMoProcessor(hf_config, server_args, _processor,
                     processor_config=json.load(open('preprocessor_config.json')))

def normalize_cfg(cfg, name):
    if cfg is None: return {}
    if isinstance(cfg, dict): return cfg
    if hasattr(cfg, 'to_dict'): return cfg.to_dict()
    raise ValueError(f'{name} must be dict-like, got {type(cfg)}')

processor_config = normalize_cfg(processor_config, 'processor_config')

def is_dict_like_config(c) -> bool:
    return c is None or isinstance(c, dict) or hasattr(c, 'to_dict')

try:
    proc = MiMoProcessor(hf_config, server_args, _processor, processor_config=cfg)
except ValueError as e:
    if 'dict-like config' in str(e) and hasattr(cfg, '__dict__'):
        proc = MiMoProcessor(hf_config, server_args, _processor, processor_config=dict(vars(cfg)))
    else:
        raise

When it happens

Trigger: Constructing MiMoProcessor with processor_config (or a sibling config) as a string path, a dataclass without to_dict, a Namespace, or a list instead of a dict/HF config object.

Common situations: Programmatic construction in tests passing config paths; HF transformers version changes where configs no longer expose to_dict; custom wrappers passing **kwargs-wrapped values.

Related errors

• cache_dit_params must be a dict, got {type(raw).__name__}.
• Invalid component residency assignment: {value!r}
• Invalid component residency assignment: {raw_selector!r}={ra
• component_attention_backends must be a dict or a comma-separ
• Component attention backend key must be a string

AI-assisted analysis of sgl-project/sglang@0132848349 (2026-08-28). Data as JSON: /api/errors/5f079917a3868f9c. Report an issue: GitHub.