metadata
dataset_info:
features:
- name: sample_id
dtype: int64
- name: input_sequence
dtype: string
- name: ground_truth_removed
dtype: string
- name: generated_removed
dtype: string
- name: full_sequence_ground_truth
dtype: string
- name: full_sequence_generated
dtype: string
- name: num_generated_tokens
dtype: int64
- name: generated_removed_audio
dtype:
audio:
sampling_rate: 44100
- name: ground_truth_removed_audio
dtype:
audio:
sampling_rate: 44100
- name: full_generated_audio
dtype:
audio:
sampling_rate: 44100
- name: full_ground_truth_audio
dtype:
audio:
sampling_rate: 44100
splits:
- name: train
num_bytes: 181261667
num_examples: 100
download_size: 141304885
dataset_size: 181261667
configs:
- config_name: default
data_files:
- split: train
path: data/train-*