Update app.py

app.py

@@ -64,30 +64,35 @@ class CompressionArtifactPredictor:
         """Predict compression quality levels for all formats."""
         img_tensor = self.preprocess(image).unsqueeze(0).to(self.device)
 
+        # SIMPLE FULL PRECISION INFERENCE - NO AUTOCAST
         with torch.no_grad():
-            with torch.cuda.amp.autocast(dtype=torch.bfloat16):
-                predictions = self.model(img_tensor).squeeze(0).cpu().float().numpy()
+            predictions = self.model(img_tensor).squeeze(0).cpu().numpy()
 
         results = {}
         for i, fmt in enumerate(self.compression_formats):
             quality_score = float(predictions[i] * 100)
             
             if quality_score >= 90:
-                category = "Excellent (Minimal artifacts)"
+                category = "Excellent"
                 color = "🟢"
+                desc = "Minimal artifacts"
             elif quality_score >= 70:
-                category = "Good (Light artifacts)"
+                category = "Good"
                 color = "🟡"
+                desc = "Light artifacts"
             elif quality_score >= 50:
-                category = "Fair (Moderate artifacts)"
+                category = "Fair"
                 color = "🟠"
+                desc = "Moderate artifacts"
             else:
-                category = "Poor (Heavy artifacts)"
+                category = "Poor"
                 color = "🔴"
+                desc = "Heavy artifacts"
             
             results[fmt] = {
                 'quality_score': round(quality_score, 1),
                 'category': category,
+                'desc': desc,
                 'accuracy': self.accuracy_scores[fmt],
                 'indicator': color
             }
@@ -101,7 +106,7 @@ def create_ui():
 
     def analyze_image(image):
         if image is None:
-            return None, "Please upload an image."
+            return "", "Please upload an image."
         
         if isinstance(image, np.ndarray):
             image = Image.fromarray(image)
@@ -109,25 +114,45 @@ def create_ui():
         image = image.convert('RGB')
         results = predictor.predict(image)
         
-        formatted_results = {}
+        # Generate HTML table for results
+        html_results = """
+        <table style='width:100%; border-collapse: collapse; font-family: inherit;'>
+            <tr style='background: #f5f5f5;'>
+                <th style='padding:12px; text-align:left; border-bottom: 2px solid #ddd;'>Format</th>
+                <th style='padding:12px; text-align:center; border-bottom: 2px solid #ddd;'>Quality</th>
+                <th style='padding:12px; text-align:center; border-bottom: 2px solid #ddd;'>Assessment</th>
+                <th style='padding:12px; text-align:center; border-bottom: 2px solid #ddd;'>Accuracy</th>
+            </tr>
+        """
+        
         for fmt, data in results.items():
-            formatted_results[f"{data['indicator']} {fmt}"] = {
-                "Predicted Quality": f"{data['quality_score']}/100",
-                "Assessment": data['category'],
-                "Model Accuracy": f"{data['accuracy']}%"
-            }
+            html_results += f"""
+                <tr style='border-bottom: 1px solid #eee;'>
+                    <td style='padding:12px; font-weight:500;'>{data['indicator']} {fmt}</td>
+                    <td style='padding:12px; text-align:center;'><strong>{data['quality_score']}/100</strong></td>
+                    <td style='padding:12px; text-align:center;'>{data['category']}<br><small style='color:#666;'>{data['desc']}</small></td>
+                    <td style='padding:12px; text-align:center;'>{data['accuracy']}%</td>
+                </tr>
+            """
+        html_results += "</table>"
         
+        # Overall summary
         avg_quality = np.mean([r['quality_score'] for r in results.values()])
         if avg_quality >= 85:
-            overall_status = "✅ **High Quality Image** - Minimal compression artifacts detected."
+            overall_status = "✅ **High Quality Image** - Minimal compression artifacts detected across all formats."
         elif avg_quality >= 65:
-            overall_status = "⚠️ **Moderate Quality** - Some compression artifacts present, but usable."
+            overall_status = "⚠️ **Moderate Quality** - Some compression artifacts present, but image remains usable."
         else:
             overall_status = "❌ **Low Quality Image** - Significant compression artifacts detected."
         
-        summary = f"### Overall Assessment\n{overall_status}\n\n**Average Quality Score: {avg_quality:.1f}/100**"
+        summary = f"""
+        ### Overall Assessment
+        {overall_status}
+        
+        **Average Quality Score: {avg_quality:.1f}/100**
+        """
         
-        return formatted_results, summary
+        return html_results, summary
 
     with gr.Blocks(
         title="AAL-Plus Image Quality Assessment",
@@ -156,9 +181,9 @@ def create_ui():
                 analyze_button = gr.Button("🔍 Analyze Image Quality", variant="primary", size="lg")
             
             with gr.Column():
-                results_output = gr.Label(
-                    label="Format-Specific Quality Scores",
-                    num_top_classes=4
+                # USE HTML COMPONENT INSTEAD OF LABEL
+                results_output = gr.HTML(
+                    label="Format-Specific Quality Scores"
                 )
                 summary_output = gr.Markdown(
                     label="Overall Assessment"
@@ -193,7 +218,6 @@ def create_ui():
 
     return demo
 
-# ==================== MAIN ====================
 if __name__ == "__main__":
     demo = create_ui()
     demo.launch()