Quantcast
[ 3 / biz / cgl / ck / diy / fa / g / ic / jp / lit / sci / tg / vr ] [ index / top / reports / report a bug ] [ 4plebs / archived.moe / rbt ]

Support us on Patreon!

/sci/ - Science & Math


View post   

[ Toggle deleted replies ]
File: 10 KB, 518x402, hh.png [View same] [iqdb] [saucenao] [google] [report]
10886890 No.10886890 [Reply] [Original] [archived.moe]

can anyone help tell me what im doing wrong here? its a simple neural net using a back propagation algorithm that i believe should work, but whenever i train it the predictions almost never change

heres my source https://github.com/frecklebars/text-color-predictor/blob/master/_js(incomplete)/tcp.js

thanks in advance

>> No.10886899

>>10886890
what this does is take 3 inputs, the RGB values of a color (from 0 to 1), feeds them forward to 5 neurons in a hidden layer and then once again in the output layer.
then it calculates the cost function and the derrivative of that with respect for each weight and bias. that gets substracted from the actual weights and biases

>> No.10886975

bump

>> No.10888467

x

>>
Name (leave empty)
Comment (leave empty)
Name
E-mail
Subject
Comment
Password [?]Password used for file deletion.
reCAPTCHA
Action