Projects are your one-stop-shop for creating a deep learning model in the Platform. A Project collects all the necessary data and processes into one place where you can manage everything from start to finish.

Let's walk through how to create a Project from scratch!

1. From the Projects page, click + Add Project along the top near the search bar. This will bring up a new Project modal.

2. Give your Project a name (try to keep it short and descriptive, e.g. "Building detector") and add more long-form details in the Description box.

3. Select a Project Type from the four options. This is a crucial step, so be sure to think carefully about which type of model you want to create! The model types are covered in “Types of CV” Article.

4. Then attach a dataset if you already have one ready for this project, or you can attach a dataset in the overview tab once your project is created!

Note: You won't be able to change the Project Type once you select it, so again: choose wisely!

5. Click Create. Your Project will be built, and you'll arrive on the Project Overview page.

The Project Overview page is a mini dashboard with some basic information about this particular Project.

1. The first thing you'll want to do when you're here is tell the Project which Dataset(s) you want to use to train your model.

2. In the Data card, search for the Dataset you'd like to attach. You can attach as many Datasets as you'd like, but make sure they're all relevant to the model you're trying to build in this Project. (For example, if you're working on a Building Detector, you'll want your attached Datasets to include buildings!)

3. Click Save. The Project will now create workflows behind the scenes for all the media in the Dataset(s) you just attached.

Your Project is successfully started! You're now ready to move on to the Annotate page, where you'll manage the creation of training data for this Project.