metadata
license: cc-by-sa-4.0
dataset_info:
features:
- name: label
dtype:
class_label:
names:
'0': Italian
'1': Scottish English
'2': Brazilian
'3': Mainstream US English
'4': German
'5': Egyptian
'6': Vietnamese
'7': Southern British English
'8': Jamaican English
'9': Catalan
'10': Latin
'11': Indian English
'12': Indonesian English
'13': European
'14': South African English
'15': Don't know
'16': Latin American
'17': Chinese
'18': Eastern European
'19': Bulgarian
'20': Irish English
'21': French
'22': Japanese
'23': Dutch
'24': Spanish
- name: audio
dtype: audio
splits:
- name: validation
num_bytes: 3643027205.336
num_examples: 9848
download_size: 3334754633
dataset_size: 3643027205.336
configs:
- config_name: default
data_files:
- split: validation
path: data/validation-*
language:
- en
Reformatted version of https://huggingface.co/datasets/edinburghcstr/edacc for audio classification
Make sure to remove the "Don't know" label:
dataset = dataset.filter(lambda example: example["label"] != 15)