Jamendo-QA: A Large-Scale Music Question Answering Dataset

2025/09/19 by Koh, Junyoung, Soo Yong Kim, Kim, Soo Yong +3
Arts and Humanities · Computer Science · #Audio and Speech Processing (eess.AS) #Diverse Musicological Studies #FOS: Computer and information sciences #FOS: Electrical engineering #Multimedia (cs.MM) #Music Technology and Sound Studies #Music and Audio Processing #Sound (cs.SD) #electronic engineering #information engineering

paper · pdf · doi:10.48550/arxiv.2509.15662

openalex publication_date 2025/09/19 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

We introduce Jamendo-QA, a large-scale dataset for Music Question Answering (Music-QA). The dataset is built on freely licensed tracks from the Jamendo platform and is automatically annotated using the Qwen-Omni model. Jamendo-QA provides question-answer pairs and captions aligned with music audio, enabling both supervised training and zero-shot evaluation. Our resource aims to fill the gap of music-specific QA datasets and foster further research in music understanding, retrieval, and generative applications. In addition to its scale, Jamendo-QA covers a diverse range of genres, instruments, and metadata attributes, allowing robust model benchmarking across varied musical contexts. We also provide detailed dataset statistics and highlight potential biases such as genre and gender imbalance to guide fair evaluation. We position Jamendo-QA as a scalable and publicly available benchmark that can facilitate future research in music understanding, multimodal modeling, and fair evaluation of music-oriented QA systems.

Citations

• MusiXQA: Advancing Visual Music Understanding in Multimodal Large Language Models
• Qwen2.5-Omni Technical Report
• JamendoMaxCaps: A Large Scale Music-caption Dataset with Imputed Metadata
• Music Understanding LLaMA: Advancing Text-to-Music Generation with Question Answering and Captioning
• AudioLDM 2: Learning Holistic Audio Generation with Self-supervised Pretraining
• LP-MusicCaps: LLM-Based Pseudo Music Captioning
• Simple and Controllable Music Generation
• MERT: Acoustic Music Understanding Model with Large-Scale Self-supervised Training
• AudioLDM: Text-to-Audio Generation with Latent Diffusion Models
• Learning to Answer Questions in Dynamic Audio-Visual Scenarios
• Training language models to follow instructions with human feedback
• HTS-AT: A Hierarchical Token-Semantic Audio Transformer for Sound Classification and Detection
• Audio Captioning using Pre-Trained Large-Scale Language Model Guided by Audio-based Similar Caption Retrieval
• Clotho: An Audio Captioning Dataset
• Sigmoid-Weighted Linear Units for Neural Network Function Approximation in Reinforcement Learning
• Sigmoid-weighted linear units for neural network function approximation in reinforcement learning
• Long Short-Term Memory

Related

• MQAD: A Large-Scale Question Answering Dataset for Training Music Large Language Models
• ArtistMus: A Globally Diverse, Artist-Centric Benchmark for Retrieval-Augmented Music Question Answering
• Music Understanding LLaMA: Advancing Text-to-Music Generation with Question Answering and Captioning
• MUST-RAG: MUSical Text Question Answering with Retrieval Augmented Generation
• JamendoMaxCaps: A Large Scale Music-caption Dataset with Imputed Metadata