sew-ft-fake-detection Open-source Audio Classification Model

Sew Ft Fake Detection

Developed by alexandreacff

This model is a fine-tuned audio classification model based on asapp/sew-mid-100k on the alexandreacff/kaggle-fake-detection dataset, designed for fake audio detection.

Audio Classification

Transformers

OtherOpen Source License:Apache-2.0 #Audio Forgery Detection #High-precision Classification #Voice Security

Downloads 58

Release Time : 5/9/2024

Model Overview

This is a fine-tuned audio classification model specifically designed for detecting fake audio. Based on the SEW-MID architecture and trained on a specific dataset, it achieves an accuracy of 74.39%.

Model Features

High Accuracy

Achieves 74.39% accuracy on the evaluation set.

Fine-tuned Model

Fine-tuned based on the SEW-MID-100k pre-trained model.

Optimized Training

Trained using the Adam optimizer and linear learning rate scheduler.

Model Capabilities

Audio Classification

Fake Audio Detection

Audio Feature Extraction

Use Cases

Security Detection

Fake Audio Identification

Detects whether the audio content has been tampered with or synthesized.

74.39% accuracy

Content Moderation

Audio Content Verification

Verifies the authenticity of audio content.

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.6344	0.9851	33	0.6395	0.6505
0.6157	2.0	67	0.6703	0.5215
0.5572	2.9851	100	0.5615	0.7131
0.4968	4.0	134	0.8149	0.6944
0.4988	4.9851	167	0.8099	0.7037
0.4756	6.0	201	0.8319	0.7103
0.4334	6.9851	234	0.7317	0.7336
0.4321	8.0	268	0.6548	0.7383
0.4436	8.9851	301	0.6232	0.7439
0.4493	9.8507	330	0.6278	0.7439

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

Sew Ft Fake Detection

Model Overview

Model Features

Model Capabilities

Use Cases

🚀 sew-ft-fake-detection

🚀 Quick Start

📚 Documentation

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

📄 License