Optimizing Large Language Models through Quantization: A Comparative Analysis of PTQ and QAT Techniques