huggingface/transformers · error · ValueError

can't derive true or false from {v} (key {k})

Error message

can't derive true or false from {v} (key {k})

What it means

ValueError from update_from_string when the target attribute is a bool but the provided value string is not one of the accepted truthy/falsy tokens (true/1/y/yes or false/0/n/no, case-insensitive). The parser refuses to guess a boolean from arbitrary text.

Source

Thrown at src/transformers/configuration_utils.py:1196

        Args:
            update_str (`str`): String with attributes that should be updated for this class.

        """

        d = dict(x.split("=") for x in update_str.split(","))
        for k, v in d.items():
            if not hasattr(self, k):
                raise ValueError(f"key {k} isn't in the original config dict")

            old_v = getattr(self, k)
            if isinstance(old_v, bool):
                if v.lower() in ["true", "1", "y", "yes"]:
                    v = True
                elif v.lower() in ["false", "0", "n", "no"]:
                    v = False
                else:
                    raise ValueError(f"can't derive true or false from {v} (key {k})")
            elif isinstance(old_v, int):
                v = int(v)
            elif isinstance(old_v, float):
                v = float(v)
            elif not isinstance(old_v, str):
                raise TypeError(
                    f"You can only update int, float, bool or string values in the config, got {v} for key {k}"
                )

            setattr(self, k, v)

    def dict_dtype_to_str(self, d: dict[str, Any]) -> None:
        """
        Checks whether the passed dictionary and its nested dicts have a *dtype* key and if it's not None,
        converts torch.dtype to a string of just the type. For example, `torch.float32` get converted into *"float32"*
        string, which can then be stored in the json format.
        """
        if d.get("dtype") is not None:

View on GitHub (pinned to a597f97485)

Solutions

1. Use one of the accepted tokens: true/false, 1/0, y/n, yes/no (any casing).
2. Strip whitespace/quotes from each value before building the update string.
3. Set booleans directly: config.scale_attn_weights = True.

Example fix

// before
config.update_from_string("scale_attn_weights=on")  # ValueError

// after
config.update_from_string("scale_attn_weights=true")

BOOL_TOKENS = {"true", "1", "y", "yes", "false", "0", "n", "no"}
for k, v in parsed_pairs.items():
    if isinstance(getattr(config, k, None), bool) and str(v).strip().lower() not in BOOL_TOKENS:
        raise ValueError(f"{k}={v!r} is not a recognized boolean token")

def is_bool_token(v: str) -> bool:
    return v.strip().strip('"\'').lower() in {"true", "1", "y", "yes", "false", "0", "n", "no"}

When it happens

Trigger: config.update_from_string("scale_attn_weights=maybe"), or passing 'True ' with whitespace, or 'on/off', or 't/f' which are not in the accepted token list.

Common situations: User-supplied CLI flags flowed into the update string with unexpected spellings; values produced by shell substitution that include quotes or spaces.

Related errors

• key {k} isn't in the original config dict
• You can only update int, float, bool or string values in the
• File not found: {audio}
• out_indices must be a list, got {type(self._out_indices)}
• out_indices must be valid indices for stage_names {self.stag

AI-assisted analysis of huggingface/transformers@a597f97485 (2026-08-14). Data as JSON: /api/errors/274ac13ca568ed5a. Report an issue: GitHub.