Why i can't use EarlyStoppingCallback and load_best_model_at_end=False

🤗Transformers
1

I want to use the trainer over 5 folds. I wanted to add EarlyStoppingCallback to the trainer function to make it stop if the training is not improving. I get this error

AssertionError: EarlyStoppingCallback requires load_best_model_at_end = True

The following is the code I used:

training_args = Seq2SeqTrainingArguments(
    output_dir="./logs",
    evaluation_strategy="epoch",
    logging_strategy="epoch",
    learning_rate=2e-5,
    per_device_train_batch_size=16,
    per_device_eval_batch_size=16,
    weight_decay=0.01,
    num_train_epochs=5,
    save_total_limit=2,
    save_strategy="epoch",
    load_best_model_at_end=True,
    predict_with_generate=True,
    fp16=False,
    push_to_hub=False,
)

for train_dataset, val_dataset in zip(train_ds, val_ds):
    trainer = Seq2SeqTrainer(
        model=model,
        args=training_args,
        train_dataset=train_dataset,
        eval_dataset=val_dataset,
        tokenizer=tokenizer,
        data_collator=data_collator,
        compute_metrics=compute_metrics,
        callbacks=[
            CombinedTensorBoardCallback,
            EarlyStoppingCallback(early_stopping_patience=3),
        ],
    )

    train_result = trainer.train()

why i can’t use EarlyStoppingCallback and load_best_model_at_end=False? I just want to save the best model at the level of each fold.
Another question: is there a way to save the best model among the 5 folds?