Custom fitness function for yolo11m-cls hyperparameter tuning

Matthew_Guay · February 10, 2025, 5:32pm

Hey all,

I’m using the model.tune hyperparameter tuning functionality to tune a custom classifier based on the yolo11m-cls.pt checkpoint. It’s all basically fine except I’d like to tweak the tuner’s fitness function.

How can I do this? I’ve searched through the documentation and e.g. the Hyperparameter Tuning Guide (Hyperparameter Tuning - Ultralytics YOLO Docs) starts off talking about the importance of identifying a fitness function, then never actually specifies how to change it from the default! Any help would be much appreciated.

Thanks,
Matt

Toxite · February 10, 2025, 11:09pm

You need to modify source code for that.

github.com/ultralytics/ultralytics

ultralytics/utils/metrics.py

9d8c2fe3c


      
          def process(self, targets, pred):
              """Target classes and predicted classes."""
              pred, targets = torch.cat(pred), torch.cat(targets)
              correct = (targets[:, None] == pred).float()
              acc = torch.stack((correct[:, 0], correct.max(1).values), dim=1)  # (top1, top5) accuracy
              self.top1, self.top5 = acc.mean(0).tolist()
          
          @property
          def fitness(self):
              """Returns mean of top-1 and top-5 accuracies as fitness score."""
              return (self.top1 + self.top5) / 2
          
          @property
          def results_dict(self):
              """Returns a dictionary with model's performance metrics and fitness score."""
              return dict(zip(self.keys + ["fitness"], [self.top1, self.top5, self.fitness]))
          
          @property
          def keys(self):
              """Returns a list of keys for the results_dict property."""
              return ["metrics/accuracy_top1", "metrics/accuracy_top5"]

Topic		Replies	Views
I Need Help with YOLOv5 Training for Custom Object Detection YOLO yolov5 , support	1	229	July 24, 2024
New Release: Ultralytics v8.3.64 Discussion releases , announcements	0	14	January 20, 2025
New Release: Ultralytics v8.2.85 Discussion releases , announcements	0	46	September 1, 2024
New Release: Ultralytics v8.3.1 Discussion releases , announcements	0	64	October 1, 2024
New Release: Ultralytics v8.2.81 Discussion releases , announcements	0	30	August 23, 2024

Custom fitness function for yolo11m-cls hyperparameter tuning

Related topics