NOVA – Neural Optimizer for Value Approximation Predict values without using many resources by utalizing neural networks.
NOVA is a cluster of neural networks designed to make predictions based of the dataset given. In this example we use the california housing dataset. It works by setting up the data correctly and feeding it inside the neural network made with pytorch. There are options for number of networks in a cluster or use the single network mode.
Currently this is just a python file and not a library (it will be soon).
You can try the demo by installing the requirements. Make sure python 3.11 is installed first.
Run pip install -r requirements.txt
After that you can run python main.py and you can test it on your own.
To run a exported model run python run.py
To intergrate this on your own please take a look at the run.py as well!
This is licensed under the MIT license and feel free to modify it!