description: Learn to simplify the logging of YOLO11 training with Comet ML. This guide covers installation, setup, real-time insights, and custom logging.
Logging key training details such as parameters, metrics, image predictions, and model checkpoints is essential in [machine learning](https://www.ultralytics.com/glossary/machine-learning-ml)—it keeps your project transparent, your progress measurable, and your results repeatable.
[Ultralytics YOLO11](https://www.ultralytics.com/) seamlessly integrates with Comet ML, efficiently capturing and optimizing every aspect of your YOLO11 [object detection](https://www.ultralytics.com/glossary/object-detection) model's training process. In this guide, we'll cover the installation process, Comet ML setup, real-time insights, custom logging, and offline usage, ensuring that your YOLO11 training is thoroughly documented and fine-tuned for outstanding results.
[Comet ML](https://www.comet.com/site/) is a platform for tracking, comparing, explaining, and optimizing machine learning models and experiments. It allows you to log metrics, parameters, media, and more during your model training and monitor your experiments through an aesthetically pleasing web interface. Comet ML helps data scientists iterate more rapidly, enhances transparency and reproducibility, and aids in the development of production models.
By combining Ultralytics YOLO11 with Comet ML, you unlock a range of benefits. These include simplified experiment management, real-time insights for quick adjustments, flexible and tailored logging options, and the ability to log experiments offline when internet access is limited. This integration empowers you to make data-driven decisions, analyze performance metrics, and achieve exceptional results.
Before diving into the usage instructions, be sure to check out the range of [YOLO11 models offered by Ultralytics](../models/index.md). This will help you choose the most appropriate model for your project requirements.
After running the training code, Comet ML will create an experiment in your Comet workspace to track the run automatically. You will then be provided with a link to view the detailed logging of your [YOLO11 model's training](../modes/train.md) process.
Comet automatically logs the following data with no additional configuration: metrics such as mAP and loss, hyperparameters, model checkpoints, interactive confusion matrix, and image [bounding box](https://www.ultralytics.com/glossary/bounding-box) predictions.
Let's dive into what you'll see on the Comet ML dashboard once your YOLO11 model begins training. The dashboard is where all the action happens, presenting a range of automatically logged information through visuals and statistics. Here's a quick tour:
The experiment panels section of the Comet ML dashboard organize and present the different runs and their metrics, such as segment mask loss, class loss, precision, and [mean average precision](https://www.ultralytics.com/glossary/mean-average-precision-map).
In the metrics section, you have the option to examine the metrics in a tabular format as well, which is displayed in a dedicated pane as illustrated here.
The confusion matrix, found in the Confusion Matrix tab, provides an interactive way to assess the model's classification [accuracy](https://www.ultralytics.com/glossary/accuracy). It details the correct and incorrect predictions, allowing you to understand the model's strengths and weaknesses.
Comet ML logs system metrics to help identify any bottlenecks in the training process. It includes metrics such as GPU utilization, GPU memory usage, CPU utilization, and RAM usage. These are essential for monitoring the efficiency of resource usage during model training.
Comet ML offers the flexibility to customize its logging behavior by setting environment variables. These configurations allow you to tailor Comet ML to your specific needs and preferences. Here are some helpful customization options:
### Logging Image Predictions
You can control the number of image predictions that Comet ML logs during your experiments. By default, Comet ML logs 100 image predictions from the validation set. However, you can change this number to better suit your requirements. For example, to log 200 image predictions, use the following code:
Comet ML allows you to specify how often batches of image predictions are logged. The `COMET_EVAL_BATCH_LOGGING_INTERVAL` environment variable controls this frequency. The default setting is 1, which logs predictions from every validation batch. You can adjust this value to log predictions at a different interval. For instance, setting it to 4 will log predictions from every fourth batch.
In some cases, you may not want to log the confusion matrix from your validation set after every [epoch](https://www.ultralytics.com/glossary/epoch). You can disable this feature by setting the `COMET_EVAL_LOG_CONFUSION_MATRIX` environment variable to "false." The confusion matrix will only be logged once, after the training is completed.
If you find yourself in a situation where internet access is limited, Comet ML provides an offline logging option. You can set the `COMET_MODE` environment variable to "offline" to enable this feature. Your experiment data will be saved locally in a directory that you can later upload to Comet ML when internet connectivity is available.
This guide has walked you through integrating Comet ML with Ultralytics' YOLO11. From installation to customization, you've learned to streamline experiment management, gain real-time insights, and adapt logging to your project's needs.
Explore [Comet ML's official documentation](https://www.comet.com/docs/v2/integrations/third-party-tools/yolov8/) for more insights on integrating with YOLO11.
Furthermore, if you're looking to dive deeper into the practical applications of YOLO11, specifically for [image segmentation](https://www.ultralytics.com/glossary/image-segmentation) tasks, this detailed guide on [fine-tuning YOLO11 with Comet ML](https://www.comet.com/site/blog/fine-tuning-yolov8-for-image-segmentation-with-comet/) offers valuable insights and step-by-step instructions to enhance your model's performance.
Additionally, to explore other exciting integrations with Ultralytics, check out the [integration guide page](../integrations/index.md), which offers a wealth of resources and information.
- **Monitor real-time insights**: Get instant feedback on your training results, allowing for quick adjustments.
- **Log extensive metrics**: Automatically capture essential metrics such as mAP, loss, hyperparameters, and model checkpoints.
- **Track experiments offline**: Log your training runs locally when internet access is unavailable.
- **Compare different training runs**: Use the interactive Comet ML dashboard to analyze and compare multiple experiments.
By leveraging these features, you can optimize your machine learning workflows for better performance and reproducibility. For more information, visit the [Comet ML integration guide](../integrations/index.md).
- **Experiment Panels**: View different runs and their metrics, including segment mask loss, class loss, and mean average [precision](https://www.ultralytics.com/glossary/precision).
- **Metrics**: Examine metrics in tabular format for detailed analysis.
- **Interactive Confusion Matrix**: Assess classification accuracy with an interactive confusion matrix.
- **System Metrics**: Monitor GPU and CPU utilization, memory usage, and other system metrics.
For a detailed overview of these features, visit the [Understanding Your Model's Performance with Comet ML Visualizations](#understanding-your-models-performance-with-comet-ml-visualizations) section.
Yes, you can enable offline logging in Comet ML by setting the `COMET_MODE` environment variable to "offline":
```python
import os
os.environ["COMET_MODE"] = "offline"
```
This feature allows you to log your experiment data locally, which can later be uploaded to Comet ML when internet connectivity is available. This is particularly useful when working in environments with limited internet access. For more details, refer to the [Offline Logging](#offline-logging) section.
