mpuneeth17/music-genre-classifier

🎵 Music Genre Classifier

A EFFICIENTNET model trained on the GTZAN dataset that classifies music audio into 10 genres:

Blues · Classical · Country · Disco · Hip-Hop · Jazz · Metal · Pop · Reggae · Rock

Usage

import torch
import librosa, numpy as np
from torchvision import transforms
from PIL import Image

# Load model
model = ...  # see app/app.py for architecture
ckpt  = torch.hub.load_state_dict_from_url(
    "https://huggingface.co/mpuneeth17/music-genre-classifier/resolve/main/best_model.pth",
    map_location="cpu"
)
model.load_state_dict(ckpt["model"])
model.eval()

Demo

👉 Try it live: Hugging Face Space

Training

• Architecture: efficientnet
• Dataset: GTZAN (10 genres × 100 songs × 10 segments = 10,000 samples)
• Input: 128×128 Mel Spectrogram images
• Optimizer: AdamW + CosineAnnealingLR
• Accuracy: ~88-92%