423A35C7/LLaMA-Factory
1
0
Fork 0
mirror of https://github.com/hiyouga/LLaMA-Factory.git synced 2025-12-27 17:20:35 +08:00
Code Issues Packages Projects Releases Wiki Activity

[v1] model loader (#9613)

Browse Source
This commit is contained in:
Yaowei Zheng
2025-12-14 11:50:52 +08:00
committed by GitHub
parent fdd24276ed
commit aeda079014
27 changed files with 449 additions and 305 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
src/llamafactory/v1/config/arg_parser.py
View File

@@ -12,7 +12,6 @@
# See the License for the specific language governing permissions and
# limitations under the License.
import json
import sys
from pathlib import Path
@@ -28,6 +27,9 @@ from .sample_args import SampleArguments
from .training_args import TrainingArguments
InputArgument = Optional[Union[dict[str, Any], list[str]]]
def validate_args(
data_args: DataArguments,
model_args: ModelArguments,
@@ -43,9 +45,7 @@ def validate_args(
raise ValueError("Quantization is not supported with deepspeed backend.")
def get_args(
args: Optional[Union[dict[str, Any], list[str]]] = None,
) -> tuple[DataArguments, ModelArguments, TrainingArguments, SampleArguments]:
def get_args(args: InputArgument = None) -> tuple[DataArguments, ModelArguments, TrainingArguments, SampleArguments]:
"""Parse arguments from command line or config file."""
parser = HfArgumentParser([DataArguments, ModelArguments, TrainingArguments, SampleArguments])
allow_extra_keys = is_env_enabled("ALLOW_EXTRA_KEYS")