Hey, what's up š Today, i'm gonna tell you how to build a simple neural network with JavaScript by your own with no AI frameworks. Let's go!
For g...
Some comments have been hidden by the post's author - find out more
For further actions, you may consider blocking this person and/or reporting abuse
I haven't understood most of it. I'll be back as soon as I get it folks
In the next article I would try to thoroughly explain how neural network works (with math examples), if you want š And you will figure it out š
Same boat but still fascinating
Hey, check out my new article about math principles of neural networks šdev.to/liashchynskyi/how-neural-ne...
Hey, your post widened my perception. JavaScript or not, neural networks are pretty cool. I am willing to research the subject
Thanks!
Post almost unrelated, but I was reading a tut about tensorflow.js and this video about tensors is veyr mathy but also very good youtube.com/watch?v=f5liqUk0ZTw
Hi, I put the code on a RunKit notebook, for anyone who wants to see it running without having to install anything on their computer: runkit.com/edo9k/tiny-rnn
For some reason I kept getting an error in the
trainfunction, the error was output_layer is not a function, so I just wrapped theoutput_layervariable in an arrow function. Now it's running.Great article, thank you.
Wow, nice article and awesome explanation!
Your derivative is wrong. First, check this math.stackexchange.com/questions/7...
And this simple network can solve AND, OR, XOR, whatever. It depends on the data and the labels. Bias is simple, why don't you just google for it? š
Luck!
This is awesome! Very easy to follow. It's always fun to try to do things without frameworks.
You got deserved star on GitHub. Very good article, nice demonstration, thank you!
ohh, thank you so much!
Understood everything til backprop but if this is the simplest Iām glad for the explanation! Thanks bro š¤Ŗ
Amazing article and really well explained. Congratulations š
And thanks for this!
thank you! š