Text this: Adaptive fine-tuning of LLMs with QLoRA adapters for enhanced understanding in cooperative multi-agent scenarios