Update app.py

app.py

@@ -2,25 +2,20 @@ from transformers import pipeline, AutoTokenizer, AutoModelForSequenceClassifica
 import gradio as gr
 import torch
 
-# Load pre-trained model and tokenizer from Hugging Face
-model_name = "distilbert-base-uncased-finetuned-sst-2-english"  # Replace with a fake news-specific model if needed
+model_name = "distilbert-base-uncased-finetuned-sst-2-english" 
 tokenizer = AutoTokenizer.from_pretrained(model_name)
 model = AutoModelForSequenceClassification.from_pretrained(model_name)
 
-# Create a pipeline for text classification
 classifier = pipeline("text-classification", model=model, tokenizer=tokenizer)
 
-# Function to detect fake or real news
 def detect_fake_news(text):
-    if not text.strip():  # Check for empty input
+    if not text.strip():  
         return "Please enter some text to analyze."
     
-    # Get prediction from the model
     result = classifier(text)
     label = result[0]['label']
     score = result[0]['score']
     
-    # Custom logic for fake/real classification
     if label == "POSITIVE" and score > 0.9:
         return f"Likely Real (Confidence: {score:.4f})"
     elif label == "NEGATIVE" and score > 0.9:
@@ -28,20 +23,15 @@ def detect_fake_news(text):
     else:
         return f"Uncertain (Confidence: {score:.4f})"
 
-# Gradio interface
 with gr.Blocks(theme=gr.themes.Default(primary_hue="blue")) as demo:
     gr.Markdown("# Cricket Fake News Detector")
     gr.Markdown("Enter a cricket-related news snippet to check if it's fake or real.")
     
-    # Input and output components
     input_text = gr.Textbox(label="Enter News Here", placeholder="e.g., India wins World Cup 2025 with one ball!")
     output_text = gr.Textbox(label="Result", interactive=False)
     
-    # Blue button for submission
     submit_btn = gr.Button("Analyze", variant="primary")
     
-    # Link the button to the detection function
     submit_btn.click(fn=detect_fake_news, inputs=input_text, outputs=output_text)
 
-# Launch the interface
 demo.launch()