This repository provides a streaming-ready version of the Yodas (Yielding Optimized Data from Any Source) dataset. It is specifically designed for training and fine-tuning large-scale Multilingual ASR models (such as OWSM v4, Mamba-based architectures, and beyond) without the need to download terabytes of raw data.
By utilizing a custom Lhotse-based loading script, this dataset allows researchers to stream audio data directly… See the full description on the dataset page:
https://huggingface.co/datasets/N02N9/yodas_owsmv4_streaming.
