RFlash/emotion-detector

Emotion Detection Model

Emotion detection model fine-tuned on the tweet_eval/emotion dataset. Detects four emotions: anger, joy, optimism, and sadness.

Model Details

• Base Model: roberta-base
• Task: Emotion Detection
• Training Dataset: tweet_eval/emotion
• Emotions: anger, joy, optimism, sadness
• Framework: PyTorch
• Language: English

Usage

from transformers import pipeline

# Load emotion classification pipeline
classifier = pipeline("text-classification", model="RFlash/emotion-detector")

# Classify text
example_result = classifier("I'm so happy today!")
print(f"Emotion: {example_result[0]['label']}, Confidence: {example_result[0]['score']:.4f}")

Emotion Label Mapping

{0: 'anger', 1: 'joy', 2: 'optimism', 3: 'sadness'}

Performance

This model was fine-tuned on the tweet_eval/emotion dataset and achieves approximately 81.5% accuracy.

Limitations

The model is specifically trained for detecting emotions in short text segments and may not perform as well on longer or more complex texts.