🔥🔥 Explore Predictions from Ultralytics models using Weights & Biases 🪄🐝

This notebook demonstrates a typical workflow of using an Ultralytics model for training, fine-tuning, and validation and performing experiment tracking, model-checkpointing, and visualization of the model's performance using Weights & Biases.

Install Dependencies

• Install Ultralytics using pip install ultralytics. In order to learn about more ways to install Ultralytics, you can check out the official docs.

• Then, you need to install the feat/ultralytics branch from W&B, which currently houses the out-of-the-box integration for Ultralytics.

Note: The Ultralytcs integration will be soon available as a fully supported feature on Weights & Biases once this pull request is merged.

Using Ultralytics with Weights & Biases

In order to use the W&B integration with Ultralytics, we need to import the wandb.yolov8.add_wandb_callback function.

Next, we initialize the YOLO model of our choice, and invoke the add_wandb_callback function on it before performing inference with the model. This would ensure that when we perform training, fine-tuning, validation, or inference, it would automatically log the experiment logs and the images overlayed with both ground-truth and the respective prediction results using the interactive overlays for computer vision tasks on W&B along with additional insights in a wandb.Table.

Sample Experiment Tracking

Epoch-wise results visualized

Next, you can check out the following notebook to learn how to perform inference and visualize predictions during training using Weights & Biases in the following notebook:

In order to learn more about using Weights & Biases with Ultralytics, you can also read the report: Supercharging Ultralytics with Weights & Biases