Starting with Neural Networks
Last updated
Last updated
In Supervisely you can do much more than just labeling, and of the most important features of the platforms in our unique ecosystem for machine learning, that unifies the best models, AI tools for analysis and model improvement, plus numerous applications built on top.
Apart from many other platforms, Supervisely is built like an OS for Computer Vision. Because of that, we made possible integration of the best machine learning models and tools on our platform. Instead of trying to put various models inside a black box, the Ecosystem of the best models and tools integrated as Supervisely Apps.
A good start for understanding how Neural Networks work in Supervisely would be Part 4 of our video course “Computer Vision with Supervisely”
Supervisely integrates fragmented state-of-the-art deep learning technologies from GitHub repositories into a user-friendly graphical interface, making them easily accessible for daily tasks.
Supervisely Apps - these are essentially forks of popular GitHub repositories, allowing users to run and interact with them via a graphical interface, overcoming fragmentation and usability issues by integrating and adding a GUI to these repositories.
The installation of a Supervisely agent on a user's machine allows GPU resources to be utilized for training and inference, providing a seamless experience of running and managing neural networks directly from a web browser.
Supervisely addresses two main issues with GitHub repositories - fragmentation (lack of connection to other CV tools) and usability bias (difficulty for less skilled users) by integrating repositories into its ecosystem and adding GUIs.
For more complex data structures like videos and medical images (DICOM), Supervisely applies the same principles of integration and GUI enhancement to mitigate even more severe fragmentation and usability issues.
Before you start training or running neural networks, you would need to connect your PC or a cloud server with GPU to Supervisely by running a simple command in your terminal. That will allow you to train neural networks and run inference right from the Supervisely web interface. You can find information on how to do this here.