# Compiling on Raspberry PI 3

**URL:** <https://community.konduit.ai/t/compiling-on-raspberry-pi-3/323>\
**Category:** ND4J\
**Created:** [March 27, 2020, 1:40pm UTC](https://community.konduit.ai/t/compiling-on-raspberry-pi-3/323 "2020-03-27T13:40:33Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![54LiNKeR](https://yyz1.discourse-cdn.com/flex035/user_avatar/community.konduit.ai/54linker/32/110_2.png) [@54LiNKeR](https://community.konduit.ai/u/54LiNKeR)\
**Post date:** [March 27, 2020, 1:40pm UTC](https://community.konduit.ai/t/compiling-on-raspberry-pi-3/323/1 "2020-03-27T13:40:33Z")

</div>

As per [here](https://github.com/eclipse/deeplearning4j/blob/master/nd4j/RaspberryPi.md) when I get to the 3rd step `build libnd4j:` please how do I fix this [error](https://gist.github.com/54LiNKeR/080f9e467e7d17ca0c57a4098cc23fe8) _“cmake  
Could not find OpenBLAS”_

---

<div class="post-metadata">

**Author:** ![treo](https://yyz1.discourse-cdn.com/flex035/user_avatar/community.konduit.ai/treo/32/47_2.png) [@treo](https://community.konduit.ai/u/treo)\
**Post date:** [March 27, 2020, 2:12pm UTC](https://community.konduit.ai/t/compiling-on-raspberry-pi-3/323/2 "2020-03-27T14:12:26Z")

</div>

Is there a reason why you are trying to build it yourself?

We do publish a linux-armhf build of the nd4j-native backend, and either that or at least a snapshot version of it _should_ work for you.

The document you’ve linked is somewhat outdated, and should probably be removed.

---

<div class="post-metadata">

**Author:** ![54LiNKeR](https://yyz1.discourse-cdn.com/flex035/user_avatar/community.konduit.ai/54linker/32/110_2.png) [@54LiNKeR](https://community.konduit.ai/u/54LiNKeR)\
**Post date:** [March 28, 2020, 7:39pm UTC](https://community.konduit.ai/t/compiling-on-raspberry-pi-3/323/3 "2020-03-28T19:39:23Z")

</div>

Hi Treo, I honestly thought this was the way to go about compiling ND4J on the raspberry pi(from doing a google search). I have a trained model from my PC that I want to deploy to the pi. could you please provide a link as per how to get started. thanks

---

<div class="post-metadata">

**Author:** ![54LiNKeR](https://yyz1.discourse-cdn.com/flex035/user_avatar/community.konduit.ai/54linker/32/110_2.png) [@54LiNKeR](https://community.konduit.ai/u/54LiNKeR)\
**Post date:** [March 31, 2020, 3:05pm UTC](https://community.konduit.ai/t/compiling-on-raspberry-pi-3/323/4 "2020-03-31T15:05:24Z")

</div>

Please how do I build ND4J on the raspberry pi automatically, say with maven or something. I have tried both **RaspberryPi.md** and **RaspberryPi2.md** from the eclipse and @[dschowta](https://github.com/dschowta)’s repo. Or are there jar files available for me to add as library in a project

---

<div class="post-metadata">

**Author:** ![treo](https://yyz1.discourse-cdn.com/flex035/user_avatar/community.konduit.ai/treo/32/47_2.png) [@treo](https://community.konduit.ai/u/treo)\
**Post date:** [March 31, 2020, 3:11pm UTC](https://community.konduit.ai/t/compiling-on-raspberry-pi-3/323/5 "2020-03-31T15:11:36Z")

</div>

Oh, sorry your last post somehow slipped by.

You should be able to use the nd4j-native backend with the linux-armhf classifier directly. But because the rpi is a bit of moving target, you might still run into issues. But we’ll have to figure them out as you run into them.

Unfortunately, I don’t have any of the more modern RPI’s on hand, so we will have to work it out together.