by Norman Koch
What is asanAI?
asanAI allows you to experiment with Machine Learning without any prior knowledge about how it works or any theory at all. You can train on your own images, or even webcam data, without ever leaving the browser. You can also use CSV files, or any arbitrary tensor data. With a click of your mouse, you are able to add or remove layers or change their settings. It’s an almost complete graphical user interface to TensorFlow.js. You don’t need to install anything – because it’s a website which runs in your browser.
What can you use asanAI for?
asanAI can be used for various purposes:
How to use asanAI
With the help of our introductionary video series, you will get an overview of the most important functionalities of asanAI.
Your models are limited to smaller sizes. asanAI can’t replace Machine Learning on a supercomputer. It is for small problems and learning the basics. You can’t branch your models, so you (for example) can either predict a bounding box for images or a category, but not both. It is currently not planned to add these features.
An optimizer is what is used to minimize the weights, biases and filters in the network, so that the loss goes down. There are different approaches for optimizers, but the default adam should work in most cases. You can chose adam and set it’s parameters in the Settings tabs. You can set other optimizers, though, if you want to check them or need to change them for your specific project.
The software is under the MIT License. This means you can edit and republish it, but you always have to link back to the original source.
This tool is based on TensorFlow.js, the Neural Network Library developed by Google. This way, everything runs in the browser, and the data you enter into asanAI will never leave your computer. Since there is no cloud behind it, your fan might spin up when you start training, because it’s your computer doing all the heavy lifting.