# Build apq8096 GPU libs with different compiler

Source: https://forum.modalai.com/topic/1134/build-apq8096-gpu-libs-with-different-compiler
Category: FAQs (https://forum.modalai.com/category/4/faqs)
Posted: 2022-07-20 09:48:25 UTC by Philemon Benner
Replies: 3 · Views: 805

## Philemon Benner · 2022-07-20 09:48:25 UTC

Hey,
i'm currently trying to build voxl-tflite-server with a newer toolchain to get it running with tensorflow 2.8.0. The tensorflow build and other dependencies worked building but i am failing with the builtin gpu libs.
![Screenshot from 2022-07-20 11-15-38.png](https://forum.modalai.com/assets/uploads/files/1658308580624-screenshot-from-2022-07-20-11-15-38.png) 
I had a look at the Voxl System Image Build, but didn't find anything regarding to these librarys. Is there a way to build them with a different compiler ?
Also are the optimization flags of the newest voxl-tflite-server compiler specific ?

## Reply by Chad Sweet (ModalAI staff) · 2022-07-20 15:05:17 UTC

Unfortunately, no. Those come from Qualcomm and there is no opportunity for change.

## Reply by Philemon Benner · 2022-07-21 12:19:50 UTC (in reply to Chad Sweet)

@Chad-Sweet Thank you for the fast answer

## Reply by Guest · 2022-07-21 22:20:19 UTC

Hey @Philemon-Benner, 

As of a few minutes ago, the latest voxl-tflite-server (v0.3.0) is now running with tensorflow v2.8.0 for apq8096. This means that you can now run any of the models that were included with qrb5165 on VOXL without dealing with these headaches. You can pull this from the dev apq8096 packages repo, or directly from here: http://voxl-packages.modalai.com/dists/apq8096/dev/binary-arm64/voxl-tflite-server_0.3.0_202207212215.ipk