metadata
dataset_info:
features:
- name: audio
dtype: audio
- name: conversation
dtype: string
- name: asr
dtype: bool
- name: prompt_lang
dtype: string
- name: answer_lang
dtype: string
- name: index
dtype: int64
splits:
- name: train
num_bytes: 433846150123.96
num_examples: 89480
download_size: 449192990384
dataset_size: 433846150123.96
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
Audio Instruct Dataset
audio instruct dataset, synthesized with (Silero tts)[https://github.com/snakers4/silero-models]