metadata
dataset_info:
features:
- name: sample_index
dtype: int64
- name: original_filename
dtype: string
- name: phonemes_hex
dtype: string
- name: phonemes_ipa
dtype: string
- name: left_audio
dtype:
audio:
sampling_rate: 16000
- name: right_audio
dtype:
audio:
sampling_rate: 16000
- name: removed_audio
dtype:
audio:
sampling_rate: 16000
- name: full_audio
dtype:
audio:
sampling_rate: 16000
- name: full_tokens_audio
dtype:
audio:
sampling_rate: 16000
- name: audio_durations
struct:
- name: left
dtype: float64
- name: right
dtype: float64
- name: removed
dtype: float64
- name: full
dtype: float64
- name: full_tokens
dtype: float64
splits:
- name: train
num_bytes: 52279396
num_examples: 100
download_size: 50159348
dataset_size: 52279396
configs:
- config_name: default
data_files:
- split: train
path: data/train-*