metadata
language: en
license: apache-2.0
library_name: transformers
tags:
- emotion-classification
- distilbert
- text-classification
- fine-tuned
datasets:
- go_emotions
Emotion Classification with DistilBERT
This model is a fine-tuned version of distilbert-base-uncased for emotion classification. It classifies text into 6 emotions:
- 0: admiration
- 1: amusement
- 2: anger
- 3: annoyance
- 4: approval
- 5: caring
Training Data
The model was fine-tuned on the Go Emotions dataset, filtered to these 6 emotion categories.
Performance
- Accuracy: 78.3%
- F1 Score: 77.9%
- Training Loss: 0.45 (from 0.93)
Usage
from transformers import pipeline
classifier = pipeline('text-classification', model='your-username/emotion-classifier-distilbert')
result = classifier('I love this amazing product!')
print(f"Emotion: {result[0]['label']}, Confidence: {result[0]['score']:.3f}")
Example Predictions
- 'I love this so much!' → admiration (confidence: ~0.85)
- 'This is so frustrating!' → anger (confidence: ~0.82)
- 'That's hilarious!' → amusement (confidence: ~0.88)
- 'This is annoying me' → annoyance (confidence: ~0.79)
- 'Great job on this!' → approval (confidence: ~0.81)
- 'I'm here to support you' → caring (confidence: ~0.83)
Training Details
- Base Model: distilbert-base-uncased
- Epochs: 3
- Batch Size: 16
- Learning Rate: 2e-5
- Dataset: Go Emotions (filtered)
Intended Use
This model is suitable for emotion analysis in text, customer feedback analysis, sentiment-aware chatbots, and social media monitoring.