Tag
This paper introduces UGTPhon, a benchmark for grapheme-to-phoneme conversion in user-generated text, and presents a compositional approach that improves performance by leveraging canonical forms.
This paper presents FastThaiG2P, a sub-millisecond Thai grapheme-to-phoneme conversion tool for TTS pipelines, achieving 0.15 ms average latency on a 27k-utterance benchmark. The authors demonstrate it by training a StyleTTS 2 Thai TTS model on a phonemized 20-hour open dataset.