metadata
dataset_info:
features:
- name: sample_id
dtype: int64
- name: noise_type
dtype: string
- name: num_chunks
dtype: int64
- name: original_duration
dtype: float32
- name: dirty_sequence
dtype: string
- name: clean_sequence
dtype: string
- name: generated_sequence
dtype: string
- name: chunk_sequences
sequence: string
- name: dirty_audio
dtype:
audio:
sampling_rate: 44100
- name: clean_audio
dtype:
audio:
sampling_rate: 44100
- name: generated_audio
dtype:
audio:
sampling_rate: 44100
splits:
- name: train
num_bytes: 91727496
num_examples: 30
download_size: 70875763
dataset_size: 91727496
configs:
- config_name: default
data_files:
- split: train
path: data/train-*