easyaligner: Forced alignment with GPU acceleration and flexible text normalization (compatible with all w2v2 models on HF Hub) [P]

easyaligner is a new, performant forced alignment library offering GPU acceleration and flexible text normalization, compatible with all w2v2 models on Hugging Face Hub. It addresses common challenges in speech-to-text preprocessing, such as handling partial transcripts, irrelevant audio, and long segments without chunking.