README.md

metadata

license: mit
task_categories:
  - audio-classification
language:
  - en
tags:
  - biology
  - dog
pretty_name: BpAE
size_categories:
  - 10K<n<100K

📦 Dataset Description

This dataset is part of the Barkopedia Challenge 🔗 https://uta-acl2.github.io/barkopedia.html

Check Training Data here:👉 ArlingtonCL2/Barkopedia_DOG_BREED_CLASSIFICATION_DATASET

This dataset contains 29,347 audio clips of dog barks labeled by dog breed. The audio samples come from 156 individual dogs across 5 dog breeds:

• shiba inu
• husky
• chihuahua
• german shepherd
• pitbull

---

📊 Per-Breed Summary

Breed	Train	Public Test	Private Test	Test (Total)	Total	Total %
chihuahua	4,427	154	231	385	4,812	16.40%
german shepherd	3,514	219	329	548	4,062	13.84%
husky	8,769	110	166	276	9,045	30.82%
pitbull	2,075	304	457	761	2,836	9.66%
shiba inu	8,110	192	290	482	8,592	29.28%
Total	26,895	980	1,472	2,452	29,347	100%

---

🐾 Labels

Each audio clip is labeled with the dog breed it belongs to.

The labels were manually verified by us.

The train_labels.csv includes:

• audio_id: Filename of the audio clip (e.g., bark_24248.wav)
• breed: Dog breed (e.g., shiba inu, pitbull)

---

🧪 Evaluation Setup

• Training set: 26,895 audio clips
• Test set: 2,452 audio clips
  • Includes:
    • 980 clips (~40%) for public leaderboard
    • 1,472 clips (~60%) for private leaderboard

---