cache

.gradio/cached_examples/15/indices.csv

@@ -0,0 +1,4 @@
+0
+1
+2
+3

.gradio/cached_examples/15/log.csv

@@ -0,0 +1,33 @@
+Detection Result,Performance Metrics,timestamp
+"{""path"": "".gradio/cached_examples/15/Detection Result/0f4a843e237f45cc683d/image.webp"", ""url"": ""/gradio_api/file=/tmp/gradio/ec172c8c5cac482f5cd3149cec4c95a810cd9bdd8b39676e261d8d2c04f6c692/image.webp"", ""size"": null, ""orig_name"": ""image.webp"", ""mime_type"": null, ""is_stream"": false, ""meta"": {""_type"": ""gradio.FileData""}}","🔄 Preprocessing:  0.09 ms
+⚙️  Inference:      94.41 ms
+📊 Postprocessing: 1.14 ms
+━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+⏱️  Total Time:     95.64 ms
+🎯 FPS:            10.45
+📈 Total Frames:   1
+",2025-12-16 10:43:38.214724
+"{""path"": "".gradio/cached_examples/15/Detection Result/7a807e887e791a89cfd1/image.webp"", ""url"": ""/gradio_api/file=/tmp/gradio/5a9f71131c8b6e743cff2071bd67257a6951a67aa00a87d95e1323704a5f2e13/image.webp"", ""size"": null, ""orig_name"": ""image.webp"", ""mime_type"": null, ""is_stream"": false, ""meta"": {""_type"": ""gradio.FileData""}}","🔄 Preprocessing:  0.06 ms
+⚙️  Inference:      98.54 ms
+📊 Postprocessing: 1.08 ms
+━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+⏱️  Total Time:     99.68 ms
+🎯 FPS:            10.03
+📈 Total Frames:   1
+",2025-12-16 10:43:38.670276
+"{""path"": "".gradio/cached_examples/15/Detection Result/d9cd32cfeb6409958eea/image.webp"", ""url"": ""/gradio_api/file=/tmp/gradio/9483b5797df1bb66cab99d79a6032e40cc372252e39d4041864c17c49b2cf05c/image.webp"", ""size"": null, ""orig_name"": ""image.webp"", ""mime_type"": null, ""is_stream"": false, ""meta"": {""_type"": ""gradio.FileData""}}","🔄 Preprocessing:  0.05 ms
+⚙️  Inference:      138.72 ms
+📊 Postprocessing: 1.73 ms
+━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+⏱️  Total Time:     140.51 ms
+🎯 FPS:            7.12
+📈 Total Frames:   1
+",2025-12-16 10:43:39.262280
+"{""path"": "".gradio/cached_examples/15/Detection Result/a4c35db755cee9e726ee/image.webp"", ""url"": ""/gradio_api/file=/tmp/gradio/3b4bac53ba462bf24dce5bb0407dd71d41b8c77d9b42d4eb616940880c432e12/image.webp"", ""size"": null, ""orig_name"": ""image.webp"", ""mime_type"": null, ""is_stream"": false, ""meta"": {""_type"": ""gradio.FileData""}}","🔄 Preprocessing:  0.03 ms
+⚙️  Inference:      57.13 ms
+📊 Postprocessing: 0.08 ms
+━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+⏱️  Total Time:     57.25 ms
+🎯 FPS:            17.46
+📈 Total Frames:   1
+",2025-12-16 10:43:39.875985

app.py

@@ -42,28 +42,39 @@ def get_available_models():
     return sorted(model_names)
 
 
-def load_model(model_name: str, device: str = "CPU"):
+def load_model(model_name: str, device: str = "CPU", confidence_threshold: float = 0.3):
     """
     Load OpenVINO model using model_api.
     
     Args:
         model_name: Name of the model (without .xml extension)
         device: Inference device (CPU, GPU, etc.)
+        confidence_threshold: Confidence threshold for predictions
     
     Returns:
         Model instance from model_api
     """
     global current_model, current_model_name
-    if current_model is not None and current_model_name == model_name:
-        return current_model
+    # Always reload model to apply new confidence threshold
     
     model_path = Path("models") / f"{model_name}.xml"
     
     if not model_path.exists():
         raise FileNotFoundError(f"Model not found: {model_path}")
     
-    print(f"Loading model: {model_name}")
-    model = Model.create_model(str(model_path), device=device)
+    print(f"Loading model: {model_name} with confidence threshold: {confidence_threshold}")
+    
+    # Set configuration based on model type
+    configuration = {}
+    if "YOLO" in model_name.upper():
+        # YOLO models use confidence_threshold and iou_threshold
+        configuration["confidence_threshold"] = confidence_threshold
+        configuration["iou_threshold"] = 0.5  # Standard IoU threshold for NMS
+    else:
+        # Other detection models typically use CONFIDENCE_THRESHOLD
+        configuration["confidence_threshold"] = confidence_threshold
+    
+    model = Model.create_model(str(model_path), device=device, configuration=configuration)
     model.get_performance_metrics().reset()
     
     current_model = model
@@ -77,7 +88,7 @@ def run_inference(
     image: np.ndarray,
     model_name: str,
     confidence_threshold: float
-) -> Tuple[Image.Image, str, str]:
+) -> Tuple[Image.Image, str]:
     """
     Perform inference and return visualized result with metrics.
     
@@ -87,21 +98,20 @@ def run_inference(
         confidence_threshold: Confidence threshold for filtering predictions
     
     Returns:
-        Tuple of (visualized_image, results_text, metrics_text)
+        Tuple of (visualized_image, metrics_text)
     """
     # Input validation
     if image is None:
-        return None, "⚠️ Please upload an image first.", ""
+        return None, "⚠️ Please upload an image first."
     
     if model_name is None or model_name == "No models available":
-        return None, "⚠️ No model selected or available.", ""
+        return None, "⚠️ No model selected or available."
     
     try:
-        model = load_model(model_name)
+        model = load_model(model_name, confidence_threshold=confidence_threshold)
         
         # Run inference
         result = model(image)
-        
         # Get performance metrics
         metrics = model.get_performance_metrics()
         inference_time = metrics.get_inference_time()
@@ -110,9 +120,7 @@ def run_inference(
         fps = metrics.get_fps()
         
         # Format metrics text
-        metrics_text = f"""⚡ Performance Metrics:
-━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
-🔄 Preprocessing:  {preprocess_time.mean()*1000:.2f} ms
+        metrics_text = f"""🔄 Preprocessing:  {preprocess_time.mean()*1000:.2f} ms
 ⚙️  Inference:      {inference_time.mean()*1000:.2f} ms
 📊 Postprocessing: {postprocess_time.mean()*1000:.2f} ms
 ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
@@ -121,17 +129,15 @@ def run_inference(
 📈 Total Frames:   {inference_time.count}
 """
         
-        # Format results based on model type
-        results_text = format_results(result, confidence_threshold)
-        
         # Visualize results using model_api's visualizer
+        print(f"Visualizing results with confidence threshold: {confidence_threshold}")
         visualized_image = visualizer.render(image, result)
         
-        return visualized_image, metrics_text, metrics_text
+        return visualized_image, metrics_text
         
     except Exception as e:
         error_msg = f"Error during inference: {str(e)}"
-        return None, error_msg, "Error: Could not compute metrics"
+        return None, error_msg
 
 
 def format_results(result, confidence_threshold: float) -> str:
@@ -206,9 +212,9 @@ def create_gradio_interface():
         print("Warning: No models found in models/ folder")
         available_models = ["No models available"]
     
-    with gr.Blocks(title="Object Detection with model_api") as demo:
-        gr.Markdown("# 🎯 Object Detection with model_api")
-        gr.Markdown("Upload an image and select a model to perform object detection using OpenVINO and model_api")
+    with gr.Blocks(title="OpenVINO with model_api") as demo:
+        gr.Markdown("# 🎯 OpenVINO with model_api")
+        gr.Markdown("Experience high-performance object detection powered by **OpenVINO™** and **model_api**. See real-time inference with detailed performance metrics.")
         
         with gr.Row():
             with gr.Column(scale=1):
@@ -244,12 +250,6 @@ def create_gradio_interface():
                     height=400
                 )
                 
-                detections_output = gr.Textbox(
-                    label="Detected Objects",
-                    lines=8,
-                    max_lines=15
-                )
-                
                 metrics_output = gr.Textbox(
                     label="Performance Metrics",
                     lines=8,
@@ -260,22 +260,22 @@ def create_gradio_interface():
         gr.Markdown("## 📸 Examples")
         gr.Examples(
             examples=[
-                ["examples/vehicles.png", "serialized" if "serialized" in available_models else available_models[0], 0.5],
-                ["examples/dog.jpg", "maskrcnn_resnet50_fpn_v2" if "maskrcnn_resnet50_fpn_v2" in available_models else available_models[0], 0.5],
-                ["examples/people-walking.png", "maskrcnn_resnet50_fpn_v2" if "maskrcnn_resnet50_fpn_v2" in available_models else available_models[0], 0.5],                
-                ["examples/zidane.jpg", "maskrcnn_resnet50_fpn_v2" if "maskrcnn_resnet50_fpn_v2" in available_models else available_models[0], 0.5],
+                ["examples/vehicles.png", "YOLO-11-N" if "YOLO-11-N" in available_models else available_models[0], 0.5],
+                ["examples/dog.jpg", "YOLO-11-S" if "YOLO-11-S" in available_models else available_models[0], 0.6],
+                ["examples/people-walking.png", "YOLO-11-M" if "YOLO-11-M" in available_models else available_models[0], 0.3],                
+                ["examples/zidane.jpg", "resnet50" if "resnet50" in available_models else available_models[0], 0.5],
             ],
             inputs=[input_image, model_dropdown, confidence_slider],
-            outputs=[output_image, detections_output, metrics_output],
+            outputs=[output_image, metrics_output],
             fn=run_inference,
-            cache_examples=False
+            cache_examples=True
         )
         
         # Connect the button to the inference function
         classify_btn.click(
             fn=run_inference,
             inputs=[input_image, model_dropdown, confidence_slider],
-            outputs=[output_image, detections_output, metrics_output]
+            outputs=[output_image, metrics_output]
         )
     
     return demo

models/YOLO-11-L.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f9256465f36a4ab2eb25b290575aecb15ae2832776790f7a4148a3daff5c1369
+size 101465149

models/YOLO-11-M.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5cca976c16d0848a09c4ed98b2703cf2020ea622bbd1b364e95b9dbaf6ab246b
+size 80468029

models/YOLO-11-S.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:39aac0b1bf65a540fdf961523cbecce68531185b344459e45cc493e63defe1b5
+size 37876237