license: mit
language:
- en
base_model:
- ehcalabres/wav2vec2-lg-xlsr-en-speech-emotion-recognition
Speech Emotion Recognition - 6-Class Classifier
This model is a fine-tuned version of ehcalabres/wav2vec2-lg-xlsr-en-speech-emotion-recognition, specifically designed to classify emotions in English speech.
🧠 Emotion Classes
The model predicts one of the following six emotions:
Happy
Angry
Disgust
Fearful
Neutral
Sad
📊 Dataset
The model was trained on the Speech Emotion Recognition dataset from Kaggle: 🔗 https://www.kaggle.com/datasets/kevinignatiuswijaya/speech-emotion-recognition-dl
🎯 Accuracy Achieved an accuracy of 84% on the test set.
🔧 Base Model Fine-tuned from the pretrained model: ehcalabres/wav2vec2-lg-xlsr-en-speech-emotion-recognition
Load model and feature extractor
model = Wav2Vec2ForSequenceClassification.from_pretrained("your-username/your-model-name") extractor = Wav2Vec2FeatureExtractor.from_pretrained("your-username/your-model-name")
Create pipeline
classifier = pipeline("audio-classification", model=model, feature_extractor=extractor)
Predict emotion
result = classifier("path/to/audio.wav") print(result)
🧪 Applications This model can be used for:
Emotion-aware virtual assistants
Mental health monitoring tools
Human-computer interaction research
Call center emotion analytics
📁 License
Ensure compliance with the licenses for both the Kaggle dataset and the pretrained model used.