File size: 515 Bytes
6bea901
 
 
 
 
071f1eb
6bea901
071f1eb
6bea901
071f1eb
6bea901
071f1eb
6bea901
071f1eb
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
---
license: mit
---
A deep learning model for audio classification, designed to detect AI-generated music.

Model architecture : EfficientNetB0 (Transfer Learning)

Framework: TensorFlow / Keras

Task: Binary Classification (AI Music vs. Real Music)

Input: Mel Spectrograms (224x224) generated from audio files.

Training Data: Trained on a balanced dataset of thousands of 
modern songs from sources like the Free Music Archive, NCS, 
and the Suno AI dataset on Kaggle, all standardized to a 
consistent bitrate.