Update app.py

app.py

@@ -1,40 +1,42 @@
 import gradio as gr
-import torch
 import shap
-import numpy as np
-import matplotlib.pyplot as plt
 from transformers import pipeline
 
-# Load sentiment model
-classifier = pipeline("sentiment-analysis")
+# Use lighter model (important for HF)
+classifier = pipeline(
+    "sentiment-analysis",
+    model="distilbert-base-uncased-finetuned-sst-2-english"
+)
 
-def predict_with_shap(text):
+# Create SHAP explainer once
+explainer = shap.Explainer(classifier)
+
+def analyze(text):
     if not text.strip():
-        return "Please enter text", None
+        return "Please enter text", ""
 
-    # Get prediction
+    # Prediction
     result = classifier(text)[0]
     label = result["label"]
     score = result["score"]
 
-    # SHAP explainer
-    explainer = shap.Explainer(classifier)
+    # SHAP values
     shap_values = explainer([text])
 
-    # Plot SHAP values
-    plt.figure()
-    shap.plots.text(shap_values[0], display=False)
+    # Convert SHAP to HTML
+    shap_html = shap.plots.text(shap_values[0], display=False)
 
-    return f"Prediction: {label} (Confidence: {score:.2f})", plt.gcf()
+    return f"Prediction: {label} (Confidence: {score:.2f})", shap_html
 
-with gr.Blocks(title="Sentiment Analysis with SHAP") as demo:
-    gr.Markdown("# Sentiment Analysis with SHAP Explanation")
+with gr.Blocks() as demo:
+    gr.Markdown("# Sentiment Analysis with SHAP")
 
     inp = gr.Textbox(lines=4, placeholder="Enter text here...")
-    out_text = gr.Textbox(label="Prediction")
-    out_plot = gr.Plot(label="SHAP Explanation")
+    prediction = gr.Textbox(label="Prediction")
+    shap_output = gr.HTML(label="SHAP Explanation")
+
     btn = gr.Button("Analyze")
 
-    btn.click(predict_with_shap, inp, [out_text, out_plot])
+    btn.click(analyze, inp, [prediction, shap_output])
 
 demo.launch()