Items: 2568
This dataset ships as a single tar at the repo root. Members preserve paths like
EARS//audio.mp3 and EARS//speaker_latent.safetensors.
See loader.py for example loading.
Attribution:
Contains audio and embeddings derived from the EARS dataset. Distributed under CC BY-NC 4.0; attribution required; commercial use is not permitted.
