Hello,
I am seeing some Github repositories that set eval_dataset with test_dataset. I was thinking that it might cause overfitting. Can you please tell me if there is a specific reason, or it is just a bug?
trainer = Trainer(
model,
args,
train_dataset=tokenized_datasets_train,
eval_dataset=tokenized_datasets_test,
tokenizer=tokenizer,
compute_metrics=compute_metrics,
data_collator = data_collator_
)