DBD-research-group/beans_watkins

---
dataset_info:
features:
- name: 'Unnamed: 0'
dtype: int64
- name: path
dtype: audio
- name: label
dtype: string
splits:
- name: train
num_bytes: 791557202.69
num_examples: 1017
- name: valid
num_bytes: 272364369.0
num_examples: 339
- name: test
num_bytes: 269221587.0
num_examples: 339
- name: train_low
num_bytes: 222944097.0
num_examples: 203
download_size: 1428752044
dataset_size: 1556087255.69
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: valid
path: data/valid-*
- split: test
path: data/test-*
- split: train_low
path: data/train_low-*
---
# Dataset Card for "beans_watkins"

## Dataset Description

**Paper:**
https://doi.org/10.1121/2.0000358

## Dataset Summary

This dataset contains annotated recordings of marine mammal sounds with splits and preprocessing like described in BEANS. It is used for **classification** tasks.

## Data Splits

| train | train_low | valid | test |
|------|-----| ------ | ------ |
| 1017 | 203 | 339 | 339 |