Face recognition model facenet with Jetson TX2

Hi,

I am working on face recognition solution and i use facenet on my host machine and it works well. So i wanted to use the same facenet (https://github.com/davidsandberg/facenet) on Jetson TX2.

I have Jetson TX2 installed with Jetpack 4.2. When i run facenet (https://github.com/davidsandberg/facenet/blob/master/contributed/export_embeddings.py) on TX2, the whole system hangs after loading the model.

Can you please tell me?

  • Is facenet architecture compatible with Jetson TX2?
  • Is there any model conversion required and how to do?

And,

  • If it is not supported, what is the best alternative to facenet on Jetson TX2? Please advise?

Hi,

May I know how do you install TensorFlow on the TX2?
Have you followed our script?
https://docs.nvidia.com/deeplearning/frameworks/install-tf-jetson-platform/index.html

Thanks.

Hi,

i have installed Tensorflow based on the below link,
https://devtalk.nvidia.com/default/topic/1038957/jetson-tx2/tensorflow-for-jetson-tx2-/

Python 3.6+JetPack4.2
sudo pip3 install --pre --extra-index-url https://developer.download.nvidia.com/compute/redist/jp/v42 tensorflow-gpu

Any updates?

Hi,

In general, incompatible should result in error rather than stuck.
Would you mind to check the memory status with tegrastats and share the log with us?

sudo tegrastats

Thanks.

Unfortunately, i don’t have hardware with me now and cannot try this issue again. Will get back to you in future if possible. Thanks.