What exactly is Integral Regularization ©?

A Python module with a mathematically advanced, proprietary algorithm for fitting a neural network to training data.  Integral Regularization © lowers validation loss for all of the many tested datasets and neural network architectures.  Runs on Linux (64-bit), Mac (64-bit), and Windows (64-bit) for Python 3.7+ and TensorFlow 2.0+, Keras, or PyTorch 1.5+.

Does Integral Regularization © work with custom code?

Yes, it easily integrates into any Python custom code including: data preprocessing, neural network architecture, regularization techniques, optimizer, loss, and so on.  A few simple lines of code are all that it takes to run.

What happens after the license expiration date?

Integral Regularization © will cease operations after the expiration date of the license.  All implementations of Integral Regularization © can use a "try/except" logical control flow which will ensure that your code will continue to function after the expiration date.  Any network trained with Integral Regularization © can be used freely outside of the licensing time period without complications or additional fees. 

What happens if I replace hardware on my computer?

Integral Regularization © will work as long as you do not reinstall operating systems.  If you want to use Integral Regularization © on both Linux and Windows on one machine, then you will need two separate licenses.

May I see the source code?

No, the source code is proprietary.


Will I get free support?

Yes, paying users will be given free support during their license period.  Please contact technicalsupport@aizia.org for technical support.