Streamline Custom Vision AI Ops

Dedicated to making computer vision accessible, we are pleased to announce our partnership with Roboflow as the official dataset management and annotation tool for Ultralytics YOLOv5.

Roboflow’s goal is to enable developers to build better computer vision models by streamlining the process of collecting and labeling your data, training your model, and using active learning to improve models even faster. This way, developers can spend more time on domain-specific problems instead of machine learning minutiae and computer vision infrastructure.

Roboflow has continuously demonstrated their support for open source projects that share our vision. Make sure you check their guide on How to Train Ultralytics YOLOv5 on Custom Dataset, released in June 2020.

Ultralytics’ mission is to make AI easy by creating accessible Vision AI tools for every developer. YOLOv5 allows developers to start training and deploying state-of-the-art vision models in their own projects.

Through a shared vision of easy AI, Roboflow and Ultralytics are combining their strengths to bring greater capability to our users, and this new partnership makes training custom models easier.

Introducing A New Partnership That Makes Training Custom Models Easier#

Ultralytics YOLOv5 has now become one of the world's most popular vision AI models due to its simplicity and accuracy in real-world use cases. Our new Roboflow integration provides an even smoother YOLOv5 training and deployment experience.

Now you can label and automatically export your custom datasets directly to Ultralytics YOLOv5 for training with Roboflow using their pip package, with a focus on enabling active learning. This way, you get to streamline your MLOps process by avoiding the time spent on computer vision minutiae.

Where to Start?#

We've made it easy to train your models on custom datasets with our new YOLOv5 Colab notebook. To get started, follow the steps below:

1. Open the YOLOv5 custom training Colab notebook
2. Run pip install roboflow to begin using your Roboflow datasets
3. Export your project from Roboflow in your desired format
4. A code snippet will be generated for custom training
5. Copy/paste the snippet from your account into the designated spot in the notebook

Roboflow + Ultralytics partnership.

You can find more useful information on the following Roboflow video tutorial.

An Exciting Time to be in the Field of Computer Vision#

The combination of Roboflow's newest features with the strengths of YOLOv5 makes for a powerful MLOps solution. We’re excited for the possibilities that Ultralytics and Roboflow are opening up through our shared vision of easy AI, and can't wait to see what our users create using our new integration.

We're looking forward to hearing about the breakthroughs that our developer community will achieve. Today, Roboflow datasets can also be imported directly into Ultralytics Platform for annotation and training.