Update app.py

app.py

@@ -1,19 +1,22 @@
 import gradio as gr
 import shap
+import torch
+import numpy as np
+import matplotlib.pyplot as plt
 from transformers import pipeline
 
-# Use lighter model (important for HF)
+# Load lightweight model
 classifier = pipeline(
     "sentiment-analysis",
     model="distilbert-base-uncased-finetuned-sst-2-english"
 )
 
-# Create SHAP explainer once
+# Create explainer
 explainer = shap.Explainer(classifier)
 
 def analyze(text):
     if not text.strip():
-        return "Please enter text", ""
+        return "Please enter text", None
 
     # Prediction
     result = classifier(text)[0]
@@ -23,20 +26,25 @@ def analyze(text):
     # SHAP values
     shap_values = explainer([text])
 
-    # Convert SHAP to HTML
-    shap_html = shap.plots.text(shap_values[0], display=False)
+    tokens = shap_values[0].data
+    values = shap_values[0].values
 
-    return f"Prediction: {label} (Confidence: {score:.2f})", shap_html
+    # Create bar plot
+    plt.figure()
+    plt.barh(tokens, values)
+    plt.xlabel("SHAP Value")
+    plt.title("Word Contribution to Sentiment")
+
+    return f"Prediction: {label} (Confidence: {score:.2f})", plt.gcf()
 
 with gr.Blocks() as demo:
     gr.Markdown("# Sentiment Analysis with SHAP")
 
     inp = gr.Textbox(lines=4, placeholder="Enter text here...")
     prediction = gr.Textbox(label="Prediction")
-    shap_output = gr.HTML(label="SHAP Explanation")
+    shap_plot = gr.Plot(label="SHAP Explanation")
 
     btn = gr.Button("Analyze")
-
-    btn.click(analyze, inp, [prediction, shap_output])
+    btn.click(analyze, inp, [prediction, shap_plot])
 
 demo.launch()