huggingface/transformers · error · ValueError
Some specified arguments are not used by the HfArgumentParse
Error message
Some specified arguments are not used by the HfArgumentParser: {remaining_args} What it means
Error "Some specified arguments are not used by the HfArgumentParser: {remaining_args}" thrown in huggingface/transformers.
Source
Thrown at src/transformers/hf_argparser.py:354
# args specified via the command line should overwrite args from files, so we add them last
args = file_args + args if args is not None else file_args + sys.argv[1:]
namespace, remaining_args = self.parse_known_args(args=args)
outputs = []
for dtype in self.dataclass_types:
keys = {f.name for f in dataclasses.fields(dtype) if f.init}
inputs = {k: v for k, v in vars(namespace).items() if k in keys}
for k in keys:
delattr(namespace, k)
obj = dtype(**inputs)
outputs.append(obj)
if len(namespace.__dict__) > 0:
# additional namespace.
outputs.append(namespace)
if return_remaining_strings:
return (*outputs, remaining_args)
else:
if remaining_args:
raise ValueError(f"Some specified arguments are not used by the HfArgumentParser: {remaining_args}")
return (*outputs,)
def parse_dict(self, args: dict[str, Any], allow_extra_keys: bool = False) -> tuple[DataClass, ...]:
"""
Alternative helper method that does not use `argparse` at all, instead uses a dict and populating the dataclass
types.
Args:
args (`dict`):
dict containing config values
allow_extra_keys (`bool`, *optional*, defaults to `False`):
Defaults to False. If False, will raise an exception if the dict contains keys that are not parsed.
Returns:
Tuple consisting of:
- the dataclass instances in the same order as they were passed to the initializer.View on GitHub (pinned to a597f97485)
Solutions
- Remove the unused CLI arguments or add matching fields to the dataclass.
- Check for typos between CLI flags and dataclass field names.
When it happens
Trigger: Raised in HfArgumentParser.parse_args_into_dataclasses when unrecognized CLI arguments remain after parsing.
Common situations: Typos in command-line flags or passing arguments for a different script to a Trainer-based entrypoint.
AI-assisted analysis of huggingface/transformers@a597f97485 (2026-08-14).
Data as JSON: /api/errors/beace41af198b4d5.
Report an issue: GitHub.