This repository is a collection of sorted neural networks. It is licensed under the BSD 3-Clause "New" or "Revised" License (see LICENSE file), so use the networks freely, but to redistribute you must make major edits and mention them. Also credits must be given to the authors.
The models given are tensorflow models. In each neural network, you will find a folder named FullModel. That folder is the tensorflow model. You will also find a readme.md, instructing how to use the model in detail, and a Test Code.py. The Test Code.py allows you to test the network. You need to have python installed to run it and also get the dependencies mentioned in the readme.md via pip. Take help from the Test Code.py to implement the neural networks in your own project.
The models were trained on an intel I7 7th generation AVX2 processor. Number of samples and epochs are different for each network and are mensioned in the readme.txt of each network. Type of neural network is also mensioned in the same. Activation function of the networks is relu.
You can see the tensorflow quickstart to learn tensorflow, pandas official website to learn pandas, and use the resources given in this repo. The DatasetMaker.py helps you to make a dataset, NNMaker.py is a neural network maker template, and once your network is made, you can take help from any TestCode.py from any network in this repo to make the TestCode.py for your own network or directly implement it in your project.