v1.0

PhoASR: Vietnamese ASR data creation pipeline

Code for the data creation pipeline in the paper "Vietnamese Automatic
Speech Recognition: A Revisit" (Findings of the Association for
Computational Linguistics: EACL 2026).

The pipeline aggregates and preprocesses noisy open-source Vietnamese
speech data into a high-quality ASR dataset with word-level timestamps,
across eight stages: validate, transcribe, score_wer, filter, punc_cap,
num2word, align, and finalize.