app.py

import AutoTokenizer, AutoModelForSequenceClassification
import torch
import gradio as gr

# Load model and tokenizer
model_name = "nateraw/bert-base-uncased-emotion"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForSequenceClassification.from_pretrained(model_name)

# Emotion labels
labels = ['sadness', 'joy', 'love', 'anger', 'fear', 'surprise']

# Prediction function
def predict_emotion(text):
    inputs = tokenizer(text, return_tensors="pt", truncation=True, padding=True)
    with torch.no_grad():
        outputs = model(**inputs)
    probs = torch.nn.functional.softmax(outputs.logits, dim=1)
    pred_class = torch.argmax(probs).item()
    emotion = labels[pred_class]
    return f"{emotion} ({probs[0][pred_class].item()*100:.2f}% confidence)"

# Gradio Interface
interface = gr.Interface(
    fn=predict_emotion,
    inputs=gr.Textbox(lines=2, placeholder="Type something here..."),
    outputs="text",
    title="BERT-based Emotion Detection",
    description="A web app that uses a fine-tuned BERT model to detect emotions from text."
)

interface.launch()