janhuenermann / neurojs

A JavaScript deep learning and reinforcement learning library.

Home Page:http://janhuenermann.com/projects/learning-to-drive

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

How to debug the Process

nellaivijay opened this issue · comments

Hi,
How can we debug the learning process and metrics of the RL Network to get the understand about the process.

Thanks
Vijay

Hi,

didn't have time to write a documentation, maybe I'll do that in the future. But for now, can you please elaborate on what exactly you want to measure. The network loss? The current reward?

If you mean the car demo, then the chart at the bottom of the page will contain this information.
(Black line = loss, lighter-green line = current reward, darker-green line = smooth reward)

Jan