YOLOv10 added

README.md

@@ -7,7 +7,7 @@ sdk: gradio
 sdk_version: 4.19.2
 app_file: app.py
 pinned: false
-license: agpl-3.0
+license: mit
 ---
 
 Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

app.py

@@ -6,7 +6,14 @@ import supervision as sv
 from inference import get_model
 
 MARKDOWN = """
-# YOLO-ARENA 🏟️
+<h1 style='text-align: center'>YOLO-ARENA 🏟️</h1>
+
+Welcome to YOLO-Arena! This demo showcases the performance of various YOLO models:
+
+- YOLOv8
+- YOLOv9
+- YOLOv10
+- YOLO-NAS
 
 Powered by Roboflow [Inference](https://github.com/roboflow/inference) and 
 [Supervision](https://github.com/roboflow/supervision).
@@ -16,77 +23,55 @@ IMAGE_EXAMPLES = [
     ['https://media.roboflow.com/dog.jpeg', 0.3]
 ]
 
-YOLO_V8_MODEL = get_model(model_id="yolov8s-640")
-YOLO_NAS_MODEL = get_model(model_id="coco/14")
+YOLO_V8_MODEL = get_model(model_id="yolov8m-640")
+YOLO_NAS_MODEL = get_model(model_id="coco/15")
 YOLO_V9_MODEL = get_model(model_id="coco/17")
+YOLO_V10_MODEL = get_model(model_id="coco/22")
 
 LABEL_ANNOTATORS = sv.LabelAnnotator(text_color=sv.Color.black())
 BOUNDING_BOX_ANNOTATORS = sv.BoundingBoxAnnotator()
 
 
-def process_image(
+def detect_and_annotate(
+    model,
     input_image: np.ndarray,
     confidence_threshold: float,
     iou_threshold: float
-) -> Tuple[np.ndarray, np.ndarray, np.ndarray]:
-    yolo_v8_result = YOLO_V8_MODEL.infer(
-        input_image,
-        confidence=confidence_threshold,
-        iou_threshold=iou_threshold
-    )[0]
-    yolo_v8_detections = sv.Detections.from_inference(yolo_v8_result)
-
-    labels = [
-        f"{class_name} {confidence:.2f}"
-        for class_name, confidence
-        in zip(yolo_v8_detections["class_name"], yolo_v8_detections.confidence)
-    ]
-
-    yolo_v8_annotated_image = input_image.copy()
-    yolo_v8_annotated_image = BOUNDING_BOX_ANNOTATORS.annotate(
-        scene=yolo_v8_annotated_image, detections=yolo_v8_detections)
-    yolo_v8_annotated_image = LABEL_ANNOTATORS.annotate(
-        scene=yolo_v8_annotated_image, detections=yolo_v8_detections, labels=labels)
-
-    yolo_nas_result = YOLO_NAS_MODEL.infer(
-        input_image,
-        confidence=confidence_threshold,
-        iou_threshold=iou_threshold
-    )[0]
-    yolo_nas_detections = sv.Detections.from_inference(yolo_nas_result)
-
-    labels = [
-        f"{class_name} {confidence:.2f}"
-        for class_name, confidence
-        in zip(yolo_nas_detections["class_name"], yolo_nas_detections.confidence)
-    ]
-
-    yolo_nas_annotated_image = input_image.copy()
-    yolo_nas_annotated_image = BOUNDING_BOX_ANNOTATORS.annotate(
-        scene=yolo_nas_annotated_image, detections=yolo_nas_detections)
-    yolo_nas_annotated_image = LABEL_ANNOTATORS.annotate(
-        scene=yolo_nas_annotated_image, detections=yolo_nas_detections, labels=labels)
-
-    yolo_v9_result = YOLO_V9_MODEL.infer(
+) -> np.ndarray:
+    result = model.infer(
         input_image,
         confidence=confidence_threshold,
         iou_threshold=iou_threshold
     )[0]
-    yolo_v9_detections = sv.Detections.from_inference(yolo_v9_result)
+    detections = sv.Detections.from_inference(result)
+    annotated_image = input_image.copy()
+    annotated_image = BOUNDING_BOX_ANNOTATORS.annotate(
+        scene=annotated_image, detections=detections)
+    annotated_image = LABEL_ANNOTATORS.annotate(
+        scene=annotated_image, detections=detections)
+    return annotated_image
 
-    labels = [
-        f"{class_name} {confidence:.2f}"
-        for class_name, confidence
-        in zip(yolo_v9_detections["class_name"], yolo_v9_detections.confidence)
-    ]
 
-    yolo_v9_annotated_image = input_image.copy()
-    yolo_v9_annotated_image = BOUNDING_BOX_ANNOTATORS.annotate(
-        scene=yolo_v9_annotated_image, detections=yolo_v9_detections)
-    yolo_v9_annotated_image = LABEL_ANNOTATORS.annotate(
-        scene=yolo_v9_annotated_image, detections=yolo_v9_detections, labels=labels)
-
-    return yolo_v8_annotated_image, yolo_nas_annotated_image, yolo_v9_annotated_image
+def process_image(
+    input_image: np.ndarray,
+    confidence_threshold: float,
+    iou_threshold: float
+) -> Tuple[np.ndarray, np.ndarray, np.ndarray, np.ndarray]:
+    yolo_v8_annotated_image = detect_and_annotate(
+        YOLO_V8_MODEL, input_image, confidence_threshold, iou_threshold)
+    yolo_nas_annotated_image = detect_and_annotate(
+        YOLO_NAS_MODEL, input_image, confidence_threshold, iou_threshold)
+    yolo_v9_annotated_image = detect_and_annotate(
+        YOLO_V9_MODEL, input_image, confidence_threshold, iou_threshold)
+    yolo_10_annotated_image = detect_and_annotate(
+        YOLO_V10_MODEL, input_image, confidence_threshold, iou_threshold)
+
+    return (
+        yolo_v8_annotated_image,
+        yolo_nas_annotated_image,
+        yolo_v9_annotated_image,
+        yolo_10_annotated_image
+    )
 
 
 confidence_threshold_component = gr.Slider(
@@ -125,21 +110,27 @@ with gr.Blocks() as demo:
     with gr.Row():
         input_image_component = gr.Image(
             type='numpy',
-            label='Input Image'
-        )
-        yolo_v8_output_image_component = gr.Image(
-            type='numpy',
-            label='YOLOv8 Output'
-        )
-    with gr.Row():
-        yolo_nas_output_image_component = gr.Image(
-            type='numpy',
-            label='YOLO-NAS Output'
-        )
-        yolo_v9_output_image_component = gr.Image(
-            type='numpy',
-            label='YOLOv9 Output'
+            label='Input'
         )
+        with gr.Column():
+            with gr.Row():
+                yolo_v8_output_image_component = gr.Image(
+                    type='numpy',
+                    label='YOLOv8m @ 640x640'
+                )
+                yolo_nas_output_image_component = gr.Image(
+                    type='numpy',
+                    label='YOLO-NAS M @ 640x640'
+                )
+            with gr.Row():
+                yolo_v9_output_image_component = gr.Image(
+                    type='numpy',
+                    label='YOLOv9c @ 640x640'
+                )
+                yolo_v10_output_image_component = gr.Image(
+                    type='numpy',
+                    label='YOLOv10m @ 640x640'
+                )
     submit_button_component = gr.Button(
         value='Submit',
         scale=1,
@@ -156,7 +147,8 @@ with gr.Blocks() as demo:
         outputs=[
             yolo_v8_output_image_component,
             yolo_nas_output_image_component,
-            yolo_v9_output_image_component
+            yolo_v9_output_image_component,
+            yolo_v10_output_image_component
         ]
     )
 
@@ -170,7 +162,8 @@ with gr.Blocks() as demo:
         outputs=[
             yolo_v8_output_image_component,
             yolo_nas_output_image_component,
-            yolo_v9_output_image_component
+            yolo_v9_output_image_component,
+            yolo_v10_output_image_component
         ]
     )