distilhubert-finetuned-gtzan Open-Source Audio Classification Model

Distilhubert Finetuned Gtzan

Developed by Gyaneshere

This model is an audio classification model fine-tuned on the GTZAN music genre classification dataset based on DistilHuBERT, with an accuracy of 84%.

Audio Classification

Transformers

Open Source License:Apache-2.0 #Music Classification #Lightweight Audio Model #High Accuracy

Downloads 5

Release Time : 2/8/2025

Model Overview

A lightweight audio classification model specifically designed for music genre recognition tasks, fine-tuned on the GTZAN dataset based on the DistilHuBERT architecture.

Model Features

Efficient and Lightweight

Based on the distilled architecture of DistilHuBERT, reducing model complexity while maintaining performance

High Accuracy

Achieved a classification accuracy of 84% on the GTZAN test set

Fast Training

Good performance can be achieved in just 10 epochs

Model Capabilities

Music Audio Classification

Genre Recognition

Audio Feature Extraction

Use Cases

Music Analysis

Automatic Music Genre Classification

Automatically classify music segments by genre

Classification accuracy of 84% for 10 music genres

Music Recommendation System

Serves as the front-end classification module of a music recommendation system

Audio Processing

Audio Content Analysis

Extract audio features for content analysis

🚀 distilhubert-finetuned-gtzan

This model is a fine - tuned version of [ntu - spml/distilhubert](https://huggingface.co/ntu - spml/distilhubert) on the GTZAN dataset, achieving an accuracy of 0.84 on the evaluation set.

🚀 Quick Start

This model is a fine - tuned version of [ntu - spml/distilhubert](https://huggingface.co/ntu - spml/distilhubert) on the GTZAN dataset. It achieves the following results on the evaluation set:

Loss: 0.5500
Accuracy: 0.84

📚 Documentation

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

🔧 Technical Details

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 5e - 05
train_batch_size: 8
eval_batch_size: 8
seed: 42
optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon = 1e - 08 and optimizer_args = No additional optimizer arguments
lr_scheduler_type: linear
lr_scheduler_warmup_ratio: 0.1
num_epochs: 10
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.0014	1.0	113	1.8735	0.56
1.2705	2.0	226	1.2514	0.63
1.0219	3.0	339	1.1399	0.65
0.7034	4.0	452	0.8793	0.69
0.5411	5.0	565	0.6780	0.81
0.4426	6.0	678	0.5764	0.84
0.2074	7.0	791	0.5895	0.81
0.1131	8.0	904	0.5772	0.8
0.1932	9.0	1017	0.5552	0.82
0.0699	10.0	1130	0.5500	0.84

Framework versions

Transformers 4.48.2
Pytorch 2.5.1+cu124
Datasets 3.2.0
Tokenizers 0.21.0

📄 License

This project is under the Apache - 2.0 license.

Property	Details
Library Name	transformers
License	apache - 2.0
Base Model	ntu - spml/distilhubert
Tags	generated_from_trainer
Datasets	marsyas/gtzan
Metrics	accuracy
Model Name	distilhubert - finetuned - gtzan
Task	Audio Classification
Dataset Name	GTZAN
Dataset Type	marsyas/gtzan
Dataset Config	all
Dataset Split	train
Dataset Args	all
Evaluation Metric	Accuracy
Metric Value	0.84

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご