Enhancing Speech Quality through the Integration of BGRU and Transformer Architectures