# TensorFlow2.16.1+nv24.08 is running very slow in Jetson AGX Orin

**URL:** <https://forums.developer.nvidia.com/t/tensorflow2-16-1-nv24-08-is-running-very-slow-in-jetson-agx-orin/320348>\
**Category:** Deep Learning (Training & Inference)\
**Tags:** computer-vision-cv, tensorflow, cuda, camera, jetson\
**Created:** [January 16, 2025, 9:53pm UTC](https://forums.developer.nvidia.com/t/tensorflow2-16-1-nv24-08-is-running-very-slow-in-jetson-agx-orin/320348 "2025-01-16T21:53:15Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![siddharthaingermany](https://developer.download.nvidia.com/images/forums/profile-default-devtalk-84.png) [@siddharthaingermany](https://forums.developer.nvidia.com/u/siddharthaingermany)\
**Post date:** [January 16, 2025, 9:53pm UTC](https://forums.developer.nvidia.com/t/tensorflow2-16-1-nv24-08-is-running-very-slow-in-jetson-agx-orin/320348/1 "2025-01-16T21:53:15Z")

</div>

I am trying to run FaceNet framework on Jetson AGX Orin 64 GB development kit using TensorFlow downloaded from [Index of /compute/redist/jp/v61](https://developer.download.nvidia.com/compute/redist/jp/v61) tensorflow==2.16.1+nv24.08. I am using Intel realsense camera. I see that speed is around 2-3 FPS. Can someone help me improving the speed !!! Usually it should be around 40 FPS on GPU.

Specifications of Orin device:  
Jetson version: Jetson 6.1  
CUDA: 12.6  
Ubuntu version: 22.04.

Here is an image of nvidia-smi that is captured when I ran this Terminal command while running a TensorFlow program. Ideally it is expected to use GPU resources.

 ![Screenshot from 2025-01-16 23-19-32](https://global.discourse-cdn.com/nvidia/original/4X/8/3/4/834a2528664a3858ea9376e29fcc6ab65c5889fe.png)

Please let me know if you need any other details. Thank you very much.