How to manage the model hyper-parameter efficiently?

Tune the model hyper-parameter is an important part with a considerable amount of works.
How to manage the different version of the model hyper-parameter efficiently?
I hope to train different version of model clearly.
Using an options file? Please give me some advice, thanks.

Hi @Derek,

Jakub from here.
It feels like Neptune would solve your problems. It helps you keep track of code, hyperparameters, data versions, summary charts (prediction distribution) and things like that. Some time ago I wrote a blog post about it, so feel free to go and read it here .

Before you ask, Neptune is open and free for non-organizations.
Read more about it on the product documentation page and check the neptune-contrib read the docs to get a better view.