|
|
--- |
|
|
language: en |
|
|
tags: |
|
|
- emotion-classification |
|
|
- text-classification |
|
|
- deberta |
|
|
license: apache-2.0 |
|
|
--- |
|
|
|
|
|
# DeBERTa Emotion Classifier |
|
|
|
|
|
This model classifies text into 5 emotions: |
|
|
- 😠 Anger |
|
|
- 😨 Fear |
|
|
- 😊 Joy |
|
|
- 😢 Sadness |
|
|
- 😲 Surprise |
|
|
|
|
|
## Usage |
|
|
```python |
|
|
from transformers import AutoTokenizer, AutoModelForSequenceClassification |
|
|
import torch |
|
|
|
|
|
model_name = "Somya26/deberta-emotion-classifier" |
|
|
tokenizer = AutoTokenizer.from_pretrained(model_name) |
|
|
model = AutoModelForSequenceClassification.from_pretrained(model_name) |
|
|
|
|
|
text = "I am so happy today!" |
|
|
inputs = tokenizer(text, return_tensors="pt") |
|
|
outputs = model(**inputs) |
|
|
probs = torch.sigmoid(outputs.logits).squeeze() |
|
|
|
|
|
emotions = ['anger', 'fear', 'joy', 'sadness', 'surprise'] |
|
|
for emotion, prob in zip(emotions, probs): |
|
|
print(f"{emotion}: {prob:.2%}") |
|
|
``` |
|
|
|
|
|
## Training |
|
|
|
|
|
Trained on emotion classification dataset using DeBERTa-v3-base. |
|
|
|
