-
Notifications
You must be signed in to change notification settings - Fork 74k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Protobuf 3.6.1 import issue with tensorflow 1.10 #21719
Comments
Thank you for your post. We noticed you have not filled out the following field in the issue template. Could you update them if they are relevant in your case, or leave them as N/A? Thanks. |
Hi, I was able to fix the issue after downgrading protobuf from 3.6.1 to 3.6.0 |
Hi,
Also, did you upgrade your tensorflow installation from a previous version, or did you install tf from scratch on your machine? |
Same here, exactly same problem after upgrading to tensorflow 1.10, also tried clean install tensorflow 1.10. protobuf 3.6.1 doesn't not work for CPU and GPU, but protobuf 3.6.0 works for both. |
**protobuf is part of tf gpu installation - meaning protobuf was installed as part of pip installation of tensorflow. |
What a nightmare getting this to work on Windows 10!
Now I can test it: |
I had exactly the same issue. I spend the entire day trying to figure out what was wrong. I installed and uninstalled python, tf, etc.. million times on my new Windows 10 machine. This morning I saw this post. I deleted protobuf 3.6.1 and installed 3.6.0. Bingo ! everything is working. Thanks a lot |
@acozzette @xfxyjwf |
@gunan No, I don't know of any backward compatibility issues between 3.6.0 and 3.6.1. @anandolee Do you have any ideas what it could be? |
This is protocolbuffers/protobuf#5046 . @TeBoring what's the solution to this issue? |
upgrade python to >=3.6.1
…On Fri, Aug 31, 2018 at 23:12 Feng Xiao ***@***.***> wrote:
This is protocolbuffers/protobuf#5046
<protocolbuffers/protobuf#5046> . @TeBoring
<https://github.com/TeBoring> what's the solution to this issue?
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#21719 (comment)>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/AE9H5Y9pGgyWmQ51UG7vzRJftmi5P8Itks5uWiVLgaJpZM4WDEPV>
.
|
Nagging Assignees @gunan, @poxvoculi: It has been 29 days with no activity and this issue has an assignee. Please update the label and/or status accordingly. |
Looking at the comments, this actually looks like a missing backwards compatibility issue in protobuf between 3.6.0 and 3.6.1 |
DLL LOAD ERROR SPECIFIED ERROR COUDN'T FOUND (SRI) C:\Users\SRIKANTH>python -c "import tensorflow" During handling of the above exception, another exception occurred: Traceback (most recent call last): Failed to load the native TensorFlow runtimE |
To add an update, this issue is not only related to tf-gpu version but also occurs for cpu version if installed with python version 3.6.0 |
I have installed tensorflow==1.5.0 and protobuf==3.6.0 and 3.4.0 alternatively but there is still an issues. /home/nkanwal/bert/lib/python3.6/site-packages/tensorflow/python/framework/dtypes.py:496: FutureWarning: Passing (type, 1) or '1type' please suggest some solution |
I din know the older versions of tensorflow were still being used. Anyway, i would suggest keeping only one version of protobuff. I think tf might be using the higher version which might be causing the issue. |
File "stylize.py", line 97, in After reading everything, Still, I am getting this error. |
Version: master/v3.6.1
Language: Python 3.6x64 on Windows 10
Steps to reproduce the behavior:
**Followed instructions mentioned on tensorflow.org's guide on installing tensorflow on Windows
(www.tensorflow.org/install/install_windows)
**protobuf is part of tf gpu installation
import tensorflow
but got an error:==>
Traceback (most recent call last):
File "<pyshell#0>", line 1, in
import tensorflow
File "C:\Users\Niraj\AppData\Local\Programs\Python\Python36\lib\site-packages\tensorflow_init_.py", line 22, in
from tensorflow.python import pywrap_tensorflow # pylint: disable=unused-import
File "C:\Users\Niraj\AppData\Local\Programs\Python\Python36\lib\site-packages\tensorflow\python_init_.py", line 52, in
from tensorflow.core.framework.graph_pb2 import *
File "C:\Users\Niraj\AppData\Local\Programs\Python\Python36\lib\site-packages\tensorflow\core\framework\graph_pb2.py", line 6, in
from google.protobuf import descriptor as _descriptor
File "C:\Users\Niraj\AppData\Local\Programs\Python\Python36\lib\site-packages\google\protobuf\descriptor.py", line 47, in
from google.protobuf.pyext import _message
ImportError: DLL load failed: The specified procedure could not be found.
Initially i thought the issue was related to CUDA or cdnn itself, but after hours of trial & error with multiple cuda & cudnn version combination, i decided to focus on the library mentioned in the error message itself.
I'm not very new to python extensions so obvious work around for me was to try earlier version of protobuf, so i tried:
pip uninstall protobuf
pip install protobuf==3.6.0
and tada!! tensorflow started working.
I'm still not sure if this was actually caused by protobuf or incorrect installations etc.
My sincere apologies if this issue is unrelated.
Please let me know if you need more details on this issue.
The text was updated successfully, but these errors were encountered: