---
dataset_info:
  features:
  - name: audio
    dtype: audio
  - name: ytid
    dtype: string
  - name: start_s
    dtype: int32
  - name: end_s
    dtype: int32
  - name: audioset_positive_labels
    dtype: string
  - name: aspect_list
    dtype: string
  - name: caption
    dtype: string
  - name: author_id
    dtype: int32
  - name: is_balanced_subset
    dtype: bool
  - name: is_audioset_eval
    dtype: bool
  splits:
  - name: train
    num_bytes: 6889418132
    num_examples: 3659
  - name: test
    num_bytes: 1697214472
    num_examples: 915
  download_size: 8336595403
  dataset_size: 8586632604
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: test
    path: data/test-*
---