Welcome to Velocity by AI4U
Use Velocity to manage the full life cycle of deep learning. Annotate and manage data sets, Convert Data Sets, continuously train and optimise custom algorithms. Velocity comes with over 40 computer vision deep learning algorithms for classification and object detection . You are able to go from idea to trained algorithm within in days. No coding needed just choose one of out of the box configured algorithms and start training.
Velocity is running on port 3000. When you access the first time http://your_ip_or_domain:3000 you will be redirected to the login page. Create an account or login as demo user. You can have limitless users on velocity. All of the artifacts of a user are not shared with other users.
You can upload media using the web interface or indicate a path where your media files are on the server. You can than mount a bucket to that path.
To train an algorithm you need annotated data. Every image needs to have a label associated to it. So that the algorithm learns to mapping between the image and the label. Labeling images is a labour intensive task. Out tool permits you to create annotations while the video is running.
You can use multiple annotation collections to create a data set. We have created a tool that permits you to do that by dragging and dropping annotation collections. In the same tool you can combine multiple data sets into one again just by dragging and dropping.
Choose one of the implemented algorithms
<iframe width="560" height="315" src="https://www.youtube.com/embed/bbxKzSOaHdo" frameborder="0" allow="accelerometer; autoplay; encrypted-media; gyroscope; picture-in-picture" allowfullscreen></iframe>