| import os | |
| import streamlit as st | |
| from transformers import pipeline | |
| auth_token = os.environ.get("TOKEN_FROM_SECRET") or True | |
| pipe = pipeline("text-classification", model="ogozcelik/bert-base-turkish-fake-news", token=auth_token) | |
| st.title("Turkish Fake News Detector") | |
| st.markdown( | |
| """ | |
| This application determines the accuracy of claims shared on social media since 2020 through Turkish Twitter texts. | |
| When you test facts that are already known to be true (such as the Earth being round), you may get misleading answers. | |
| """ | |
| ) | |
| st.write("\n\n") | |
| def print_result(user_input): | |
| result = pipe(user_input) | |
| veracity = result[0]["label"] | |
| confidence = result[0]["score"] | |
| return veracity, confidence | |
| st.markdown("Enter a tweet or short statement to detect its truthfulness:") | |
| user_input = st.text_input("") | |
| if st.button('Predict Truthfulness'): | |
| with st.spinner('Processing...'): | |
| veracity, confidence = print_result(user_input) | |
| st.success(f"Truthfulness: {veracity}") | |
| st.write(f"Confidence: The model is {(confidence*100):.2f}% sure about the result.") |