About adding the original optimizer

This question is to the effect that there is a good way to add the original optimizer to the Chainer.

We are currently testing the new optimizer on CNN, which is made only with the regular Python library and Cupy, without using the framework.

Currently, we have tested MNIST and Cifar-10 to ensure that the new optimizer is faster than major algorithms such as Adam.

In the end, I would like to configure ResNet and test it with ILSVRC 2012 data, but I wondered if I could put the algorithm in a proven framework and test it.The Cupy was very easy to use and stable, so I'm trying to figure out if I can make it happen with a Chainer.

I looked at the optimizer code for the Chainer in GitHub, but it seems to have a complex library mechanism as well as optimization algorithms, so I can't figure out how to do it.

If you have any suggestions on how to add the original optimizer to the Chainer, please let me know.Thank you for your cooperation.