423A35C7/LLaMA-Factory
1
0
Fork 0
mirror of https://github.com/hiyouga/LLaMA-Factory.git synced 2025-12-16 11:50:35 +08:00
Code Issues Packages Projects Releases Wiki Activity

add swanlab

Browse Source
This commit is contained in:
hiyouga
2024-12-19 07:12:31 +00:00
parent 6ccd64ecd9
commit 96f8f103e5
6 changed files with 60 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
src/llamafactory/train/sft/trainer.py
View File

@@ -29,7 +29,7 @@ from ...extras import logging
from ...extras.constants import IGNORE_INDEX
from ...extras.packages import is_transformers_version_equal_to_4_46, is_transformers_version_greater_than
from ..callbacks import PissaConvertCallback, SaveProcessorCallback
from ..trainer_utils import create_custom_optimizer, create_custom_scheduler
from ..trainer_utils import create_custom_optimizer, create_custom_scheduler, get_swanlab_callback
if TYPE_CHECKING:
@@ -71,6 +71,9 @@ class CustomSeq2SeqTrainer(Seq2SeqTrainer):
self.accelerator.clip_grad_norm_ = MethodType(clip_grad_norm_old_version, self.accelerator)
self.add_callback(BAdamCallback)
if finetuning_args.use_swanlab:
self.add_callback(get_swanlab_callback(finetuning_args))
@override
def create_optimizer(self) -> "torch.optim.Optimizer":
if self.optimizer is None: