fix mrope

Former-commit-id: 55bee1d333549ca19858b3f5c1b7b86926e5fb09

src/llamafactory/train/sft/trainer.py

@@ -27,7 +27,7 @@ from typing_extensions import override
 
 from ...extras import logging
 from ...extras.constants import IGNORE_INDEX
-from ...extras.packages import is_transformers_version_equal_to_4_46
+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
 
@@ -51,6 +51,9 @@ class CustomSeq2SeqTrainer(Seq2SeqTrainer):
     def __init__(
         self, finetuning_args: "FinetuningArguments", processor: Optional["ProcessorMixin"], **kwargs
     ) -> None:
+        if is_transformers_version_greater_than("4.46"):
+            kwargs["processing_class"] = kwargs.pop("tokenizer")
+
         super().__init__(**kwargs)
         self.finetuning_args = finetuning_args
 

src/llamafactory/train/sft/workflow.py

@@ -56,6 +56,7 @@ def run_sft(
 
     data_collator = SFTDataCollatorWith4DAttentionMask(
         template=template,
+        model=model,
         pad_to_multiple_of=8 if training_args.do_train else None,  # for shift short attention
         label_pad_token_id=IGNORE_INDEX if data_args.ignore_pad_token_for_loss else tokenizer.pad_token_id,
         block_diag_attn=model_args.block_diag_attn,