423A35C7/LLaMA-Factory
1
0
Fork 0
mirror of https://github.com/hiyouga/LLaMA-Factory.git synced 2025-08-04 04:32:50 +08:00
Code Issues Packages Projects Releases Wiki Activity

fix #3316

Browse Source 
Former-commit-id: c9a477322df82fecdb268ed385e3e0c376c0baeb
This commit is contained in:
hiyouga 2024-04-17 22:54:34 +08:00
parent f9c859e97b
commit 0170ef83a6
1 changed files with 6 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
src/llmtuner/model/utils.py
View File

@ -1,3 +1,4 @@
import inspect
from enum import Enum, unique
from functools import partial
from typing import TYPE_CHECKING, Any, Dict, List, Optional
@ -129,7 +130,11 @@ def gradient_checkpointing_enable(
return gradient_checkpointing_func(func, *args, **kwargs)
self._set_gradient_checkpointing(enable=True, gradient_checkpointing_func=custom_gradient_checkpointing_func)
if "value" in inspect.signature(self._set_gradient_checkpointing).parameters: # old GC format
self.apply(partial(self._set_gradient_checkpointing, value=True))
logger.warning("You are using the old GC format, some features (e.g. BAdam) will be invalid.")
else:
self._set_gradient_checkpointing(enable=True, gradient_checkpointing_func=custom_gradient_checkpointing_func)
def load_valuehead_params(path_or_repo_id: str, model_args: "ModelArguments") -> Dict[str, torch.Tensor]: