Dialect Identification Using Resource-Efficient Fine-Tuning Approaches