transformers version requirement
#18
by
GenmX9610 - opened
It seems that if we want to load model with transformers for simple validation, the suggested version 4.57.1+ is not compatible.
Since LossKwargs has been removed from transformers 4.54.0, and the modeling_deepseek.py file in the repo depends the class.
Maybe 4.53.3 is better for pure transfomers loading.
You can insert code:
from typing import Optional, TypedDict
class LossKwargs(TypedDict, total=False):
"""
Keyword arguments to be passed to the loss function
Attributes:
num_items_in_batch (`int`, *optional*):
Number of items in the batch. It is recommended to pass it when
you are doing gradient accumulation.
"""
num_items_in_batch: Optional[int]