README.md

metadata

dataset_info:
  - config_name: large
    features:
      - name: audio
        dtype: audio
      - name: class
        dtype: string
      - name: class_id
        dtype: int64
    splits:
      - name: train
        num_bytes: 1866384595
        num_examples: 37634
    download_size: 1863057968
    dataset_size: 1866384595
  - config_name: small
    features:
      - name: audio
        dtype: audio
      - name: class
        dtype: string
      - name: class_id
        dtype: int64
    splits:
      - name: train
        num_bytes: 49598479
        num_examples: 1001
    download_size: 49517177
    dataset_size: 49598479
configs:
  - config_name: large
    data_files:
      - split: train
        path: large/train-*
  - config_name: small
    data_files:
      - split: train
        path: small/train-*
task_categories:
  - audio-classification
tags:
  - audio
  - text
size_categories:
  - 10K<n<100K

The urban-noise dataset consists of audio samples representing urban noise environments. It is designed for tasks such as noise classification, audio tagging, or machine learning applications in sound analysis. The dataset includes two configurations, large and small, with varying sizes of data.

Example Usage

from datasets import load_dataset

# Load the large configuration
large_dataset = load_dataset("Sunbird/urban-noise", "large")

# Load the small configuration
small_dataset = load_dataset("Sunbird/urban-noise", "small")