Datasets:
metadata
annotations_creators:
- no-annotation
language:
- en
license:
- mit
task_categories:
- audio-classification
- anomaly-detection
tags:
- dcase
- audio
- anomaly-detection
pretty_name: DCASE 2025 Task 2 Dev
dataset_info:
features:
- name: file_id
dtype: string
- name: split
dtype: string
- name: machine_type
dtype:
class_label:
names:
'0': bearing
'1': fan
'2': gearbox
'3': slider
'4': ToyCar
'5': ToyTrain
'6': valve
- name: label
dtype:
class_label:
names:
'0': normal
'1': anomaly
'2': unknown
- name: domain
dtype:
class_label:
names:
'0': source
'1': target
'2': unknown
- name: section
dtype: string
- name: d1p
dtype:
class_label:
names:
'0': 'n'
'1': none
'2': pro
'3': v1pat
- name: d1v
dtype:
class_label:
names:
'0': 0
'1': 4
'2': 5
'3': A
'4': B
'5': C
'6': D
'7': E
'8': F
'9': X
'10': 'Y'
'11': Z
'12': none
'13': 'y'
- name: d2p
dtype:
class_label:
names:
'0': none
'1': v2pat
- name: d2v
dtype:
class_label:
names:
'0': 0
'1': 5
'2': none
- name: d3p
dtype:
class_label:
names:
'0': none
- name: d3v
dtype:
class_label:
names:
'0': none
- name: audio_path
dtype: string
- name: audio
dtype: audio
DCASE 2025 Task 2 - Development Dataset
Attributes d1v, d2v, d3v are encoded as ClassLabels.
Usage
from datasets import load_dataset
dataset = load_dataset('HTill/dcase2025_task2_dev', trust_remote_code=True)