Starting a project on Roboflow, importing images, using the annotation workflow, and adding images from Roboflow Universe

Starting a project on Roboflow, importing images, using the annotation workflow, adding images from Roboflow Universe, and more: Getting Started with Roboflow

Topics Covered:

Creating Your Project

What are annotation batches? Uploading images (working with annotation batches, and our annotation workflow) Launch: Collaborative Annotations

Getting Images from Roboflow Universe to add to a project

Uploading images to your annotation batch (fresh unannotated images, already annotated images, video files)

Using the annotation tool

Completing an annotation batch, and generating versions of your dataset for model training

Uploading the images from Roboflow Universe to the project

Navigating annotation batches, and uploading new images

Generating versions: quick primer on preprocessing and augmentation – for more: Image Transformations - Roboflow

Primer on Dataset Health Check – insights into your dataset like: total number of annotations, heatmap of annotations, average image sizes, number of labels for each class (class balance)

Training your model

Using Label Assist

Training a custom model (https://models.roboflow.com)

Roboflow Train results: model inference/deployment options and running tests Inference - Object Detection - Roboflow

2 Likes

Is there a similar tutorial for how to get started with a project from scratch using the python SDK?

Hi @VictorZuanazzi

Usually, the Roboflow Python SDK is used to assist with training a model on one of our many notebooks. Our notebooks also have a ton of resources, including blog guides and YouTube videos, as well as the GitHub repo and paper, all listed so that would be a great place to start.

If you’d like to learn how to use the Roboflow SDK specifically, here is the roboflow-python GitHub and the page on our docs.

Thanks for the reply. The python SDK documentation is a bit bare, but the alternative resources are helping.