app.py

# -*- coding: utf-8 -*-
"""app.ipynb

Automatically generated by Colab.

Original file is located at
    https://colab.research.google.com/drive/1CbDOX8PDJB6ZyLZiLMXbPyr6k7dvrs20
"""

import gradio as gr
import torch
from transformers import AutoModelForSequenceClassification, AutoTokenizer

# Load the model and tokenizer
model_name = "qarib/bert-base-qarib"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForSequenceClassification.from_pretrained(model_name, num_labels=2)

# Preprocessing function
def light_preprocess(text):
    text = text.replace("@USER", "").replace("RT", "").strip()
    return text

# Prediction function
def predict_offensive(text):
    preprocessed_text = light_preprocess(text)
    inputs = tokenizer(preprocessed_text, return_tensors="pt", truncation=True, padding=True)
    with torch.no_grad():
        outputs = model(**inputs)
    logits = outputs.logits
    predicted_class = torch.argmax(logits, dim=1).item()
    return "Offensive" if predicted_class == 1 else "Not Offensive"

# Create the Gradio interface
iface = gr.Interface(
    fn=predict_offensive,
    inputs=gr.Textbox(lines=2, placeholder="Enter text here..."),
    outputs="text",
    title="Offensive Language Detection",
    description="Enter a text to check if it's offensive or not.",
)

# Launch the interface
iface.launch(share=True)