diff --git a/boxmot/motion/cmc/sof.py b/boxmot/motion/cmc/sof.py
index 49b388503b..89aaf8c603 100644
--- a/boxmot/motion/cmc/sof.py
+++ b/boxmot/motion/cmc/sof.py
@@ -146,6 +146,9 @@ def apply(self, img, dets):
         except Exception as e:
             LOGGER.warning(f'Affine matrix could not be generated: {e}')
             return H
+        finally:
+            if H is None:
+                return np.eye(2, 3)
 
         if self.draw_optical_flow:
             self.warped_img = cv2.warpAffine(self.prev_img, H, (w, h), flags=cv2.INTER_LINEAR)
diff --git a/examples/track.py b/examples/track.py
index 0fdb9a46f0..28b2e34427 100644
--- a/examples/track.py
+++ b/examples/track.py
@@ -4,13 +4,13 @@
 
 import torch
 
-from boxmot import DeepOCSORT
+from boxmot.tracker_zoo import create_tracker
 from boxmot.utils import EXAMPLES, ROOT, WEIGHTS
 from boxmot.utils.checks import TestRequirements
 from examples.detectors import get_yolo_inferer
 
 __tr = TestRequirements()
-__tr.check_packages(('git+https://github.com/mikel-brostrom/ultralytics.git',))  # install
+__tr.check_packages(('ultralytics @ git+https://github.com/mikel-brostrom/ultralytics.git', ))  # install
 
 from ultralytics import YOLO
 from ultralytics.yolo.data.utils import VID_FORMATS
@@ -27,18 +27,20 @@ def on_predict_start(predictor, persist=False):
         predictor (object): The predictor object to initialize trackers for.
         persist (bool, optional): Whether to persist the trackers if they already exist. Defaults to False.
     """
-    predictor.args.tracking_config = \
+    tracking_config = \
         ROOT /\
         'boxmot' /\
         'configs' /\
-        'deepocsort.yaml'
+        (predictor.custom_args.tracking_method + '.yaml')
     trackers = []
     for i in range(predictor.dataset.bs):
-        tracker = DeepOCSORT(
-            model_weights=Path(WEIGHTS / 'osnet_x0_25_msmt17.pt'),
-            device=predictor.device,
-            fp16=False,
-            per_class=False
+        tracker = create_tracker(
+            predictor.custom_args.tracking_method,
+            tracking_config,
+            predictor.custom_args.reid_model,
+            predictor.device,
+            predictor.custom_args.half,
+            predictor.custom_args.per_class
         )
         trackers.append(tracker)
 
@@ -52,7 +54,6 @@ def run(args):
     yolo = YOLO(
         'yolov8n.pt',
     )
-    print(yolo.__dict__.keys())
 
     results = yolo.track(
         source=args.source,
@@ -83,9 +84,11 @@ def run(args):
     yolo.predictor.args.name = args.name
     yolo.predictor.args.exist_ok = args.exist_ok
     yolo.predictor.args.classes = args.classes
+    yolo.predictor.custom_args = args
 
     for frame_idx, r in enumerate(results):
-        if len(r.boxes.data) != 0:
+
+        if r.boxes.data.shape[1] == 7:
 
             if yolo.predictor.source_type.webcam or args.source.endswith(VID_FORMATS):
                 p = yolo.predictor.save_dir / 'mot' / (args.source + '.txt')
@@ -99,7 +102,6 @@ def run(args):
                     yolo.predictor.mot_txt_path,
                     r,
                     frame_idx,
-                    frame_idx,
                 )
 
             if args.save_id_crops:
diff --git a/examples/utils.py b/examples/utils.py
index c6304e7c8c..8e1242468a 100644
--- a/examples/utils.py
+++ b/examples/utils.py
@@ -3,12 +3,11 @@
 from ultralytics.yolo.utils import ops
 
 
-def write_mot_results(txt_path, results, frame_idx, i):
+def write_mot_results(txt_path, results, frame_idx):
     nr_dets = len(results.boxes)
     frame_idx = torch.full((1, 1), frame_idx + 1)
     frame_idx = frame_idx.repeat(nr_dets, 1)
     dont_care = torch.full((nr_dets, 1), -1)
-    i = torch.full((nr_dets, 1), i)
     mot = torch.cat([
         frame_idx,
         results.boxes.id.unsqueeze(1).to('cpu'),