Convert frozen graph to TensorRT engine

Robotics & Edge Computing Jetson & Embedded Systems Jetson TX2
1

Hi,

I get following error while converting to plan file from frozen graph using convert_plan.py
However I have python and tensorrt installed. Am I missing missing other configuration!!

Traceback (most recent call last):
File “scripts/convert_plan.py”, line 6, in
import uff
ImportError: No module named uff

Thanks,
Sachin

2

Hi,

Where do you run this python script?

If you run it on x86 Linux machine, please recheck the installation of uff package with this document:
http://docs.nvidia.com/deeplearning/sdk/tensorrt-install-guide/index.html#installing-debian

sudo apt-get install uff-converter-tf

If you run it on Jetson, check this page for information:

Thanks.

3

Hi

I encountered the same error of ImportError: No module named uff.
As I converted frozen graph to TensorRT engine with Jetpack3.3 without manaully install UFF.
Did I miss any or should I manually install Tensor RT 4 and install uff?

The thing is I converted frozen graph successfully with Jetpack 3.2 by following
the link https://github.com/NVIDIA-Jetson/tf_to_trt_image_classification.

Thank you,

4

Hi,

I still got an error as I converted frozen graph to TensorRT engine.
And also manually install TensorRT 4 and uff.
The error message is as the below.

But I converted mobilenet_v1_0p25_128.plan as I manaully installed TensorRT 3 and uff.
is it correct way to do ?

error message :
nvidia@tegra-ubuntu:~/tf_to_trt_image_classification$ python scripts/convert_plan.py data/frozen_graphs/mobilenet_v1_0p25_128.pb data/plans/mobilenet_v1_0p25_128.plan input 128 128 MobilenetV1/Logits/SpatialSqueeze 1 0 float
Traceback (most recent call last):
File “scripts/convert_plan.py”, line 71, in
data_type
File “scripts/convert_plan.py”, line 22, in frozenToPlan
text=False,
File “/usr/local/lib/python2.7/dist-packages/uff/converters/tensorflow/conversion_helpers.py”, line 149, in from_tensorflow_frozen_model
return from_tensorflow(graphdef, output_nodes, preprocessor, **kwargs)
File “/usr/local/lib/python2.7/dist-packages/uff/converters/tensorflow/conversion_helpers.py”, line 41, in from_tensorflow
https://docs.nvidia.com/deeplearning/sdk/tensorrt-api/#python and click on the ‘TensoRT Python API’ link"“”.format(err))
ImportError: ERROR: Failed to import module (No module named graphsurgeon)
Please make sure you have graphsurgeon installed.
For installation instructions, see:
https://docs.nvidia.com/deeplearning/sdk/tensorrt-api/#python and click on the ‘TensoRT Python API’ link

Thank you,

5

Hi,

Duplicate to the topic 1045099:
Please check that topic for the following udpate:
[url]https://devtalk.nvidia.com/default/topic/1045099/jetson-tx2/failed-to-convert-frozen-graph-to-tensorrt-engine/[/url]

Thanks.