AST-ASVspoof2019 Open-source Synthetic Speech Detection Model

AST ASVspoof2019 Synthetic Voice Detection

Developed by MattyB95

This model is a fine-tuned version based on MIT/ast-finetuned-audioset-10-10-0.4593, designed for detecting synthetic voices.

Audio Classification

Transformers

EnglishOpen Source License:MIT #Voice Anti-spoofing Detection #High-precision Audio Classification #ASVspoof2019 Adaptation

Downloads 18

Release Time : 12/22/2023

Model Overview

The model is primarily used for audio classification tasks, especially for detecting synthetic voices in the ASVspoof2019 dataset.

Model Features

High Precision

Achieves a precision of 0.874 on the validation set, effectively reducing false positives.

Balanced Performance

F1 score of 0.769, achieving a good balance between precision and recall.

AST Architecture

Utilizes the Audio Spectrogram Transformer architecture, suitable for processing audio spectrogram data.

Model Capabilities

Synthetic Voice Detection

Audio Classification

Voice Authenticity Verification

Use Cases

Security Verification

Voice Biometric Anti-spoofing

Detects synthetic voice attacks in voice biometric systems

Can help systems identify and block 0.685 of synthetic voice attacks

Content Moderation

Synthetic Audio Content Identification

Identifies synthetic audio on social media and content platforms

Property	Details
Model Type	AST - ASVspoof2019 - Synthetic - Voice - Detection
Base Model	MIT/ast - finetuned - audioset - 10 - 10 - 0.4593
Tags	generated_from_trainer
Datasets	audiofolder, LanceaKing/asvspoof2019
Metrics	accuracy, f1, precision, recall

Task	Dataset	Split	Metric	Value
Audio Classification	audiofolder	validation	Accuracy	0.6294477539848655
Audio Classification	audiofolder	validation	F1	0.7685655387400071
Audio Classification	audiofolder	validation	Precision	0.8743850817984212
Audio Classification	audiofolder	validation	Recall	0.6855938284894152

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

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1	Precision	Recall
	1.0
	2.0
	3.0

AST ASVspoof2019 Synthetic Voice Detection

Model Overview

Model Features

Model Capabilities

Use Cases

🚀 AST-ASVspoof2019-Synthetic-Voice-Detection

🚀 Quick Start

📚 Documentation

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

📄 License

📦 Model Information

📊 Evaluation Results