dynann/emotion-classification

Multimodal Emotion Speech Recognition

Model Description

This model performs emotion recognition from speech using a multimodal approach, utilizing:

• Audio Model: Wav2Vec2 Base

Dataset

• Dataset Name: stapesai/ssi-speech-emotion-recognition

Evaluation Results

Classification Report

              precision    recall  f1-score   support

         ANG       0.97      0.93      0.95        30
         CAL       0.00      0.00      0.00         0
         DIS       0.95      0.90      0.92        20
         FEA       0.76      0.70      0.73        27
         HAP       0.87      0.82      0.84        33
         NEU       0.96      0.96      0.96        25
         SAD       0.73      1.00      0.84        19
         SUR       0.88      0.78      0.82         9

    accuracy                           0.87       163
   macro avg       0.76      0.76      0.76       163
weighted avg       0.88      0.87      0.87       163

Overall Accuracy: 87%