CT-GateNet: Music Genre Classification Results on GTZAN, FMA-SMALL, and FMA-Medium

A research paper and associated data presenting the Convolutional Neural Network-Gated Transformer Network (CT-GateNet) for music genre classification. The method was tested on three public datasets, achieving accuracies of 98.72% on GTZAN, 89.42% on FMA-SMALL, and 69.07% on FMA-Medium. The dataset, authored by Yunyan Ma and last updated in April 2026, includes a data augmentation strategy based on a denoising diffusion probabilistic model.

Use Cases

• Benchmarking new music genre classification models based on the reported performance metrics.
• Studying the application of denoising diffusion probabilistic models for audio data augmentation.
• Analyzing the effectiveness of hybrid CNN-Transformer architectures with gated attention mechanisms for discriminative feature learning.
• Investigating model generalization capabilities across datasets of different scales like GTZAN, FMA-SMALL, and FMA-Medium.

Strengths

• Classification accuracies are explicitly reported: 98.72% on GTZAN, 89.42% on FMA-SMALL, and 69.07% on FMA-Medium.
• The method is validated on three established public music genre datasets.
• The dataset is licensed under CC-BY-4.0, permitting open sharing and adaptation.

Limitations

• The dataset is very small at 5.5 KB, suggesting it likely contains summary results or metadata rather than raw audio data.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment for certain analyses.

Provenance

Source

figshare, authored by Yunyan Ma.

Collection Method

Experiments conducted on three public music genre datasets: GTZAN, FMA-SMALL, and FMA-Medium.

Time Range

The temporal coverage of the underlying music datasets is not specified.

Freshness

Last updated 2026-04-09 17:37:56; freshness should be verified.

Geography

The geographic origin of the music data is not specified.