Automatic Speech Recognition (ASR) models fine-tuned and optimized from pre-trained architectures for enhanced perfomance on specialized datasets.