Improve yolov8 config (#988)

Co-authored-by: fatih cagatay akyon <[email protected]>

sahi/models/yolov8.py

@@ -59,18 +59,15 @@ def perform_inference(self, image: np.ndarray):
         if self.model is None:
             raise ValueError("Model is not loaded, load it by calling .load_model()")
 
-        if self.image_size is not None:  # ADDED IMAGE SIZE OPTION FOR YOLOV8 MODELS:
-            prediction_result = self.model(
-                image[:, :, ::-1], imgsz=self.image_size, verbose=False, device=self.device
-            )  # YOLOv8 expects numpy arrays to have BGR
-        else:
-            prediction_result = self.model(
-                image[:, :, ::-1], verbose=False, device=self.device
-            )  # YOLOv8 expects numpy arrays to have BGR
-
-        prediction_result = [
-            result.boxes.data[result.boxes.data[:, 4] >= self.confidence_threshold] for result in prediction_result
-        ]
+        kwargs = {"cfg": self.config_path, "verbose": False, "conf": self.confidence_threshold, "device": self.device}
+
+        if self.image_size is not None:
+            kwargs = {"imgsz": self.image_size, **kwargs}
+
+        prediction_result = self.model(image[:, :, ::-1], **kwargs)  # YOLOv8 expects numpy arrays to have BGR
+
+        # We do not filter results again as confidence threshold is already applied above
+        prediction_result = [result.boxes.data for result in prediction_result]
 
         self._original_predictions = prediction_result