Why does the neural network stop training due to the gradient?

Question

Why does the neural network stop training due to the gradient?

I train the neural network, and it stopped training due to the gradient stop condition. From what I see, the gradient of 8.14e ^{-0.6 is} greater than the minimum gradient of 1e ^-0.5 why did it stop? Is it because the gradient did not improve, so it didn't last long?

I am very new to neural networks (and using MATLAB nntool), so any help / explanation would be greatly appreciated.

Neural Network Training Performance

+3

matlab neural-network

Gjhix Apr 27 '12 at 20:19

source share

1 answer

alfa · Accepted Answer · 2012-04-27T20:39:36+0000

This is not a neural network problem, this is a problem of understanding floating point representations:

<p> 8.14e-06 = 8.14 × 10 ^ -6 = 0.00000814 <0.00001 = 1.0x10 ^ -5 = 1e-05

Why does the neural network stop training due to the gradient?

More articles: