Skip to content

Dataset · chikingsley

L2-ARCTIC accented English speech with phonetic labels

chikingsley/l2-arctic-release-v5.0

Speech corpus of 26,867 audio readings by 24 non-native English speakers with manual phoneme labeling for ASR and pronunciation analysis.

What's inside

Contains 26,867 audio readings (27.1 hours total) from 24 non-native English speakers. The data includes manual phoneme-level annotations, text prompts, and speaker-specific archives in zip format. This release mirrors the fifth version of the well-known L2-ARCTIC corpus.

What you can do with it

  • Training automated pronunciation assessment and speech error detection systems
  • Fine-tuning ASR models for improved recognition of heavy English accents
  • Developing voice conversion and accent modification models
Tasks
Automatic speech recognition, Audio classification
Languages
English
Size
10K–100K rows
Downloads this month
22.4K
Likes
1

How to load it

from datasets import load_dataset

ds = load_dataset("chikingsley/l2-arctic-release-v5.0")