# Switch Net 4 - Neural Network

**URL:** https://discourse.processing.org/t/switch-net-4-neural-network/33220
**Category:** Gallery
**Created:** [November 1, 2021, 10:37am UTC](https://discourse.processing.org/t/switch-net-4-neural-network/33220 "2021-11-01T10:37:31Z")
**Posts on this page:** 1
**Showing post:** 3

<div class="post-metadata">

### Author: ![seanc4s](https://avatars.discourse-cdn.com/v4/letter/s/e19adc/32.png) [@seanc4s](https://discourse.processing.org/u/seanc4s)
#### Post date: [November 1, 2021, 9:51pm UTC](https://discourse.processing.org/t/switch-net-4-neural-network/33220/3 "2021-11-01T21:51:14Z")

</div>

Basically you are knitting together multiple tiny width 4 neural network layers using the fast Walsh Hadamard transform as a connectionist device, to make a deep neural network.  
The mini-layer width doesn’t have to be 4, it could be 2, 8, 1, 16 or whatever.  
I used Beyond ReLU for the neural layers, but you could use any non-linear neural layer type or even just some parametric function type.  
In sum total the tiny layers use far less parameters and are far faster than a conventional layer would.  
[Beyond ReLU](https://discourse.processing.org/t/relu-is-half-a-cookie/32134)

---

_[View the full topic](https://discourse.processing.org/t/switch-net-4-neural-network/33220)._