JetPack 3.3 fails to install uff package

Robotics & Edge Computing Jetson & Embedded Systems Jetson TX2
1

Hi,
I have Jetson TX2 platform with JetPack 3.3.
I tried to install uff package with following commands
sudo apt-get update
sudo apt-get install uff-converter-tf

Reading package lists… Done
Building dependency tree
Reading state information… Done
E: Unable to locate package uff-converter-tf

In previous version (JetPack 3.2) I have successfully installed uff package by following command but
this compatible for TensorRT-3.

sudo pip3 install TensorRT-3.0.4/uff/uff-0.2.0-py2.py3-none-any.whl

Is there any way to install uff package on Jetson TX2 with JetPack 3.3?

Regards,
Alex

2

Hi,
same thing over here. I can’t parse my TF model for Jetson TX2 because uff.from_tensorflow is cannot be imported?

3

Hi,

Uff library is included in our x86 package. Please download it on our website:
https://developer.nvidia.com/nvidia-tensorrt-4x-download

And install via these commands:

$ tar -xzf TensorRT-4.0.1.6.Ubuntu-16.04.4.x86_64-gnu.cuda-9.0.cudnn7.1.tar.gz
$ sudo pip install TensorRT-4.0.1.6/uff/uff-0.4.0-py2.py3-none-any.whl

Thanks.

4

Hi,

I’m having the same problem. I’m working on Jetson Tx2, Jetpack 3.3 and I need uff converter C++ API. I have tried to use tensorrt_4.0.2.0-1+cuda9.0_arm64.deb which is downloaded in Jetpack installation but it is not solved. How can be installed?

Thanks

5

The UFF workflow consists of two components.

  1. UFF exporter (Python, distributed for desktop only)
  2. UFF parser (C++, distributed for desktop and Jetson)

However, because the exporter is a pure python package, you can actually install the UFF package distributed for desktop on the Jetson. This is done as mentioned above.