Astie Finetuned On Shemo
This model is a fine-tuned version of the AST model on the shEMO dataset, primarily used for speech emotion recognition tasks.
Downloads 24
Release Time : 4/27/2023
Model Overview
ASTie is a speech emotion recognition model based on the Audio Spectrogram Transformer (AST) architecture, fine-tuned on the shEMO dataset, capable of identifying emotional features in speech.
Model Features
High Accuracy
Achieves 81.67% accuracy on the evaluation set.
Efficient Inference
Can process approximately 8 samples per second.
AST-based Architecture
Utilizes the advanced Audio Spectrogram Transformer architecture.
Model Capabilities
Speech Emotion Recognition
Audio Classification
Use Cases
Emotion Analysis
Customer Service Emotion Monitoring
Real-time analysis of customer emotions in service calls.
81.67% recognition accuracy.
Psychological State Assessment
Evaluates the speaker's psychological state through voice analysis.
Featured Recommended AI Models
Š 2025AIbase