Forums - issues installing SNPE

2 posts / 0 new
Last post
issues installing SNPE
mvtech
Join Date: 6 Feb 20
Posts: 2
Posted: Sun, 2020-02-23 01:29

Hi,

 

I'm trying to install the x86 part of SNPE (v1.35), following the installation instructions that come with it, without success.

When installing on Ubuntu 14.04 I'm having trouble finding suitable TensorFlow / ONNX supported installations.

When installing on Ubuntu 16.04 it seems a lot of parts are missing, for example:

 

(venv) /venv/snpe-1.35.0.698$ snpe-tensorflow-to-dlc -h

Traceback (most recent call last):

  File "/venv/snpe-1.35.0.698/bin/x86_64-linux-clang/snpe-tensorflow-to-dlc", line 20, in <module>

    from snpe.converters.tensorflow.tf_to_ir import TFConverter

ImportError: No module named snpe.converters.tensorflow.tf_to_ir

after fix:

 

(venv) /venv/snpe-1.35.0.698$ snpe-tensorflow-to-dlc -h

Traceback (most recent call last):

  File "/venv/snpe-1.35.0.698/bin/x86_64-linux-clang/snpe-tensorflow-to-dlc", line 20, in <module>

    from snpe.converters.tensorflow.tf_to_ir import TFConverter

  File "/venv/snpe-1.35.0.698/lib/python/snpe/converters/tensorflow/tf_to_ir.py", line 24, in <module>

    from snpe.converters.tensorflow.loader import ModelLoader

  File "/venv/snpe-1.35.0.698/lib/python/snpe/converters/tensorflow/loader.py", line 20, in <module>

    import tensorflow.contrib

ImportError: No module named contrib

after fix:

 

(venv) /venv/snpe-1.35.0.698$ snpe-tensorflow-to-dlc --input_network tf-model.pb --input_dim input "1,299,299,3" --out_node "OutNode" --output_path tf-model.dlc --allow_unconsumed_nodes

Traceback (most recent call last):

  File "/venv/snpe-1.35.0.698/bin/x86_64-linux-clang/snpe-tensorflow-to-dlc", line 46, in <module>

    main()

  File "/venv/snpe-1.35.0.698/bin/x86_64-linux-clang/snpe-tensorflow-to-dlc", line 29, in main

    session = tf.Session(config=tf.ConfigProto(allow_soft_placement=True))

AttributeError: 'module' object has no attribute 'Session'

 

and this of course goes on and on and on, finding myself debugging the SNPE code.... this can't be right.

What am I missing?

  • Up0
  • Down0
gesqdn-forum
Join Date: 4 Nov 18
Posts: 184
Posted: Tue, 2020-05-19 23:31

Hi mvtech,

Kindly check the TensorFlow version installed using the below command,

$ pyhton3 -m pip list

 TF2 version is not fully supported yet.  So, it is suggested to use the TF 1.13.1 or TF 1.15.
We run the latest NPE 1.38 with the Ubuntu 16 having TF 1.13.1 without any issues.

  • Up0
  • Down0
or Register

Opinions expressed in the content posted here are the personal opinions of the original authors, and do not necessarily reflect those of Qualcomm Incorporated or its subsidiaries (“Qualcomm”). The content is provided for informational purposes only and is not meant to be an endorsement or representation by Qualcomm or any other party. This site may also provide links or references to non-Qualcomm sites and resources. Qualcomm makes no representations, warranties, or other commitments whatsoever about any non-Qualcomm sites or third-party resources that may be referenced, accessible from, or linked to this site.