fix examples

app.py

@@ -1,38 +1,41 @@
 import gradio as gr
 from transformers import pipeline
+from datasets import load_dataset
+from PIL import Image
+import requests
+import io
 
-# Load a model trained on car brands/makes
-# Using a model from the Hugging Face hub that's good at vehicle classification
-car_classifier = pipeline("image-classification", 
-                         model="dima806/car_models_image_detection")
+# Load your custom dataset
+def load_my_dataset():
+    try:
+        dataset = load_dataset("Avmromanov/tripoexamples")
+        return dataset
+    except Exception as e:
+        print(f"Error loading dataset: {e}")
+        return None
 
-# Alternative models you can try:
-# - "nickmuchi/vit-finetuned-cars-biased"
-# - "saltacc/anime-ai-detect"
-# - Or use a general object detection + custom logic
+# Load car classification model
+car_classifier = pipeline("image-classification", 
+                         model="dima806/car_brand_classification")
 
 def identify_car(image):
     if image is None:
         return "Please upload an image of a car"
     
     try:
-        # Convert to RGB if necessary
         if image.mode != 'RGB':
             image = image.convert('RGB')
         
-        # Get predictions
         predictions = car_classifier(image)
         
-        # Format results
         result_text = "🚗 Car Identification Results:\n\n"
+        top_5 = predictions[:5]
         
-        top_5 = predictions[:5]  # Get top 5 predictions
         for i, pred in enumerate(top_5, 1):
             label = pred['label'].replace('_', ' ').title()
             confidence = pred['score']
             result_text += f"{i}. {label}: {confidence:.2%}\n"
         
-        # Try to extract additional information
         result_text += f"\n🔍 Most likely: **{top_5[0]['label'].replace('_', ' ').title()}** " \
                       f"(confidence: {top_5[0]['score']:.2%})"
         
@@ -41,123 +44,67 @@ def identify_car(image):
     except Exception as e:
         return f"Error processing image: {str(e)}"
 
-def enhanced_car_analysis(image):
-    """Enhanced version with more detailed analysis"""
-    if image is None:
-        return "Please upload an image of a car"
+def get_dataset_examples(dataset, num_examples=3):
+    """Extract example images from the dataset"""
+    examples = []
+    
+    if dataset is None:
+        return examples
     
     try:
-        # Get basic predictions
-        predictions = car_classifier(image)
-        
-        # Create detailed analysis
-        analysis = "🚗 **Detailed Car Analysis**\n\n"
-        
-        # Top prediction
-        top_pred = predictions[0]
-        brand = top_pred['label'].replace('_', ' ').title()
-        confidence = top_pred['score']
-        
-        analysis += f"**Primary Identification:**\n"
-        analysis += f"• Brand: {brand}\n"
-        analysis += f"• Confidence: {confidence:.2%}\n\n"
-        
-        # Top 3 alternatives
-        analysis += "**Alternative Possibilities:**\n"
-        for i, pred in enumerate(predictions[1:4], 2):
-            alt_brand = pred['label'].replace('_', ' ').title()
-            alt_conf = pred['score']
-            analysis += f"{i}. {alt_brand} ({alt_conf:.2%})\n"
-        
-        # Confidence level interpretation
-        if confidence > 0.8:
-            analysis += f"\n✅ **High confidence** - Very likely a {brand}"
-        elif confidence > 0.6:
-            analysis += f"\n⚠️ **Moderate confidence** - Probably a {brand}"
-        else:
-            analysis += f"\n❓ **Low confidence** - Could be a {brand} or similar brand"
-        
-        # Tips for better identification
-        analysis += "\n\n**💡 Tips for better identification:**"
-        analysis += "\n• Use clear, front/side views of the car"
-        analysis += "\n• Ensure good lighting and focus"
-        analysis += "\n• Avoid images with multiple cars"
-        analysis += "\n• Crop close to the car for better accuracy"
-        
-        return analysis
-        
+        # Adjust this based on your dataset structure
+        train_data = dataset['train']
+        
+        for i in range(min(num_examples, len(train_data))):
+            example = train_data[i]
+            
+            # The structure depends on your dataset - adjust accordingly
+            if 'image' in example:
+                # If images are stored in the dataset
+                examples.append(example['image'])
+            elif 'url' in example:
+                # If URLs are provided
+                examples.append(example['url'])
+            elif 'path' in example:
+                # If file paths are provided
+                examples.append(example['path'])
+                
     except Exception as e:
-        return f"Error processing image: {str(e)}"
-
-# Create a more advanced interface with tabs
-with gr.Blocks(theme=gr.themes.Soft()) as demo:
-    gr.Markdown("# 🚗 Car Brand & Model Identifier")
-    gr.Markdown("Upload a photo of a car to identify its brand and model!")
+        print(f"Error extracting examples: {e}")
     
-    with gr.Tab("Quick Identification"):
-        gr.Markdown("### Fast car brand identification")
-        with gr.Row():
-            with gr.Column():
-                image_input = gr.Image(label="Upload Car Photo", type="pil")
-                quick_btn = gr.Button("Identify Car", variant="primary")
-            with gr.Column():
-                quick_output = gr.Textbox(label="Identification Results", lines=8)
+    return examples
+
+# Load your dataset
+my_dataset = load_my_dataset()
+dataset_examples = get_dataset_examples(my_dataset, num_examples=4)
+
+# Create the interface
+with gr.Blocks() as demo:
+    gr.Markdown("# 🚗 Car Identifier with My Dataset")
+    gr.Markdown("Using examples from: **Avmromanov/tripoexamples**")
     
-    with gr.Tab("Detailed Analysis"):
-        gr.Markdown("### Comprehensive car analysis")
-        with gr.Row():
-            with gr.Column():
-                detail_input = gr.Image(label="Upload Car Photo", type="pil")
-                detail_btn = gr.Button("Analyze Car", variant="primary")
-            with gr.Column():
-                detail_output = gr.Textbox(label="Detailed Analysis", lines=15)
+    with gr.Row():
+        with gr.Column():
+            image_input = gr.Image(label="Upload Car Photo", type="pil")
+            identify_btn = gr.Button("Identify Car", variant="primary")
+        with gr.Column():
+            output_text = gr.Textbox(label="Results", lines=10)
     
-    with gr.Tab("About"):
-        gr.Markdown("""
-        ## About This Car Identifier
-        
-        **How it works:**
-        - Uses a deep learning model trained on car images
-        - Identifies car brands/makes from photographs
-        - Provides confidence scores for each prediction
-        
-        **Best practices:**
-        - Use clear, well-lit photos
-        - Front or side views work best
-        - Avoid blurry or distant shots
-        - Single car images yield better results
-        
-        **Supported brands include:** Toyota, Honda, Ford, BMW, Mercedes, Audi, 
-        Volkswagen, Nissan, Chevrolet, Hyundai, and many more!
-        """)
-    
-    # Examples section
-    gr.Markdown("### Example Car Images")
+
+    gr.Markdown(f"### Dataset Examples (showing {len(dataset_examples)} samples)")
     gr.Examples(
-        examples=[
-            ["https://cdn.pixabay.com/photo/2015/01/19/13/51/car-604019_1280.jpg"],  # Audi
-            ["https://cdn.pixabay.com/photo/2012/11/02/13/02/car-63930_1280.jpg"],   # Ford Mustang
-            ["https://cdn.pixabay.com/photo/2015/09/02/12/25/bmw-918408_1280.jpg"],  # BMW
-            ["https://cdn.pixabay.com/photo/2014/09/07/22/34/car-tire-438467_1280.jpg"]  # Mercedes
-        ],
+        examples=dataset_examples,
         inputs=image_input,
-        outputs=quick_output,
+        outputs=output_text,
         fn=identify_car,
         cache_examples=True
     )
     
-    # Connect buttons
-    quick_btn.click(identify_car, inputs=image_input, outputs=quick_output)
-    detail_btn.click(enhanced_car_analysis, inputs=detail_input, outputs=detail_output)
-
-# For better performance on Hugging Face Spaces
-def load_model():
-    # This will cache the model loading
-    return pipeline("image-classification", 
-                   model="dima806/car_brand_classification")
-
-# Pre-load the model when the app starts
-car_classifier = load_model()
+    gr.Markdown("""
+    **Dataset Information:**
+    - Name: tripoexamples
+    - Author: Avmromanov
+    - Type: Car images for identification
+    """)
 
-if __name__ == "__main__":
-    demo.launch(share=True)
+demo.launch()