[ 3 / biz / cgl / ck / diy / fa / ic / jp / lit / sci / vr / vt ] [ index / top / reports ] [ become a patron ] [ status ]
2023-11: Warosu is now out of extended maintenance.

/sci/ - Science & Math

Search:


View post   

>> No.16057420 [View]
File: 16 KB, 766x419, matrix-representation-of-dot-product-1626103121.png [View same] [iqdb] [saucenao] [google]
16057420

I'm programming a MNIST number recognizer using neural networks. I have a general understanding of the forward propagation and dot products where when I take a 10 number matrix and multiply it with another 10 number matrix I will receive a SINGLE summation.
Now after forward propagation is BACKWARDS PROPAGATION. This is whats really tripping me up right now. The python program I am looking at does a dot product between two 10 number matrices and returns a 10x10 matrix. Mathematically how the fuck is this happening!?

Navigation
View posts[+24][+48][+96]