[Misc] add handler HF_TOKEN is emptry string (#20369)

Signed-off-by: rongfu.leng <rongfu.leng@daocloud.io>
This commit is contained in:
rongfu.leng 2025-07-03 00:14:31 +08:00 committed by GitHub
parent d265414dbc
commit 139508a418
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194

View File

@ -56,6 +56,22 @@ MISTRAL_CONFIG_NAME = "params.json"
logger = init_logger(__name__) logger = init_logger(__name__)
def _get_hf_token() -> Optional[str]:
"""
Get the HuggingFace token from environment variable.
Returns None if the token is not set, is an empty string,
or contains only whitespace.
This follows the same pattern as huggingface_hub library which
treats empty string tokens as None to avoid authentication errors.
"""
token = os.getenv('HF_TOKEN')
if token and token.strip():
return token
return None
_CONFIG_REGISTRY_OVERRIDE_HF: dict[str, type[PretrainedConfig]] = { _CONFIG_REGISTRY_OVERRIDE_HF: dict[str, type[PretrainedConfig]] = {
"mllama": MllamaConfig "mllama": MllamaConfig
} }
@ -195,7 +211,7 @@ def file_or_path_exists(model: Union[str, Path], config_name: str,
return file_exists(str(model), return file_exists(str(model),
config_name, config_name,
revision=revision, revision=revision,
token=os.getenv('HF_TOKEN', None)) token=_get_hf_token())
def patch_rope_scaling(config: PretrainedConfig) -> None: def patch_rope_scaling(config: PretrainedConfig) -> None:
@ -322,7 +338,7 @@ def get_config(
model, model,
revision=revision, revision=revision,
code_revision=code_revision, code_revision=code_revision,
token=os.getenv('HF_TOKEN', None), token=_get_hf_token(),
**kwargs, **kwargs,
) )
@ -334,7 +350,7 @@ def get_config(
model, model,
revision=revision, revision=revision,
code_revision=code_revision, code_revision=code_revision,
token=os.getenv('HF_TOKEN', None), token=_get_hf_token(),
**kwargs, **kwargs,
) )
else: else:
@ -344,7 +360,7 @@ def get_config(
trust_remote_code=trust_remote_code, trust_remote_code=trust_remote_code,
revision=revision, revision=revision,
code_revision=code_revision, code_revision=code_revision,
token=os.getenv('HF_TOKEN', None), token=_get_hf_token(),
**kwargs, **kwargs,
) )
except ValueError as e: except ValueError as e:
@ -571,7 +587,7 @@ def get_sentence_transformer_tokenizer_config(model: str,
# If model is on HuggingfaceHub, get the repo files # If model is on HuggingfaceHub, get the repo files
repo_files = list_repo_files(model, repo_files = list_repo_files(model,
revision=revision, revision=revision,
token=os.getenv('HF_TOKEN', None)) token=_get_hf_token())
except Exception: except Exception:
repo_files = [] repo_files = []
@ -862,7 +878,7 @@ def try_get_safetensors_metadata(
get_safetensors_metadata, get_safetensors_metadata,
model, model,
revision=revision, revision=revision,
token=os.getenv('HF_TOKEN', None), token=_get_hf_token(),
) )
try: try: