Manivannan Murugavel
1 min readOct 7, 2018

--

Hi Filip Kočica,

Your training is going good and i did highlighted the avg loss. Your wait till the avg loss decreasing 0.06. If your training avg loss not decreasing you can cut the process.

Important think is only every 10k iterations saved your weights. So you can kill the process when the iterations comes 10k,20k,30k and etc…

--

--

Manivannan Murugavel
Manivannan Murugavel

Written by Manivannan Murugavel

Artificial Intelligence and Data Science

Responses (1)