app.py

import os
import gradio as gr
from transformers import pipeline
import huggingface_hub



# Load the pre-trained model
classifier = pipeline("text-classification", model="ICILS/xlm-r-icils-ilo", device=0)

# Define the prediction function
def classify_text(text):
    """
    Classify the input text into occupational categories using a pre-trained model.
    
    Args:
        text (str): Job description text.
    
    Returns:
        tuple: (label, score) - The classification label and the associated confidence score.
    """
    result = classifier(text)[0]
    label = result['label']
    score = result['score']
    return label, score

# Create the Gradio interface
demo = gr.Interface(
    fn=classify_text,
    inputs=gr.Textbox(lines=2, label="Job Description Text", placeholder="Enter a job description..."),
    outputs=[gr.Textbox(label="ISCO-08 Label"), gr.Number(label="Score")],
    title="XLM-R ISCO Classification",
    description=(
        "Classify job descriptions into occupational categories using a pre-trained XLM-R-ISCO model "
        "from Hugging Face Spaces."
    ),
)

# Run the Gradio app
if __name__ == "__main__":
    demo.launch()