simcop2387/AutoGPTQ
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages 1 Wiki Activity

add new args of save_quantized method to push_to_hub method

Browse source
This commit is contained in:
student686 2023-10-07 13:59:53 +08:00
parent fc1184e7bc
commit 22af50bab0
1 changed files with 3 additions and 1 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

4
auto_gptq/modeling/_base.py
View file

@ -468,6 +468,8 @@ class BaseGPTQForCausalLM(nn.Module, PushToHubMixin):
private: Optional[bool] = None, private: Optional[bool] = None,
token: Optional[Union[bool, str]] = None, token: Optional[Union[bool, str]] = None,
create_pr: Optional[bool] = False, create_pr: Optional[bool] = False,
max_shard_size: str = "10GB",
model_base_name: Optional[str] = None
) -> str: ) -> str:
""" """
Upload the model to the Hugging Face Hub. Upload the model to the Hugging Face Hub.
@ -505,7 +507,7 @@ class BaseGPTQForCausalLM(nn.Module, PushToHubMixin):
if save_dir is not None: if save_dir is not None:
logger.info(f"Saving model to {save_dir}") logger.info(f"Saving model to {save_dir}")
self.save_quantized(save_dir, use_safetensors, safetensors_metadata) self.save_quantized(save_dir, use_safetensors, safetensors_metadata, max_shard_size, model_base_name)
repo_url = create_repo( repo_url = create_repo(
repo_id=repo_id, token=token, private=private, exist_ok=True, repo_type="model" repo_id=repo_id, token=token, private=private, exist_ok=True, repo_type="model"