TypeError: run() got an unexpected keyword argument 'argv'

[kopxiong]

When I followed the TensorFlow Mechanics 101 tutorial python fully_connected_feed.py in examples/tutorials/mnist directory, I came across this problem:

Traceback (most recent call last):
  File "fully_connected_feed.py", line 277, in <module>
    tf.app.run(main=main, argv=[sys.argv[0]] + unparsed)
TypeError: run() got an unexpected keyword argument 'argv'

I am using tensorflow 0.11 version in Anaconda environment.

[girving]

The argv argument was added recently: 6812d46. You are comparing a new version of a tutorial with an old version of tensorflow.

[marteiro]

I'm having the same problem. And i'm really using new example with old TF because the pip installation is not placing the examples files, so i've manually copied from sources and placed inside the pip installation.

I'm using the fallowing:
https://storage.googleapis.com/tensorflow/linux/cpu/tensorflow-0.11.0-cp27-none-linux_x86_64.whl

what should i use to get TF with examples?

thanks in advance.

[xinchenxp]

I've meet the same problem with lastest TF, does sb know how to fix this problem？Is there a new tutorial docuement that i can follow, thanks a lot !

Traceback (most recent call last):
File "convolutional.py", line 339, in
tf.app.run(main=main, argv=[sys.argv[0]] + unparsed)
TypeError: run() got an unexpected keyword argument 'argv'

[hokjoung]

Checkout the r0.11 branch instead of the master branch.
Use the tutorial code in that branch.

[xinchenxp]

Thanks a lot, i git the r0.11 banch eventually, and it works !

[marteiro]

Thanks, it worked. =D

[ttytt1978]

How to checkout the r0.11branch? my system is Ubuntu 16.04 LTS. Thanks.

[marteiro]

Use "-b r0.11" with Git clone!

[ttytt1978]

@marteiro Thanks,it worked.
Can solve this problem by Update Tensorflow to new version? How to update Tensorflow to new version?

[ttytt1978]

@girving
Can solve this problem by Update Tensorflow to new version? How to update Tensorflow to new version?

[marteiro]

@ttytt1978
If your are not contributing to the master branch, just use the r0.11. Not much else to hope for.

[loretoparisi]

I have the same error running the example in the learn folder:

on latest version 0.12:

https://github.com/tensorflow/tensorflow/blob/r0.12/tensorflow/examples/learn/text_classification.py

Traceback (most recent call last):
  File "text_classification.py", line 136, in <module>
    tf.app.run(main=main, argv=[sys.argv[0]] + unparsed)
TypeError: run() got an unexpected keyword argument 'argv'

[girving]

@loretoparisi You are not running 0.12 tensorflow: https://github.com/tensorflow/tensorflow/blob/r0.12/tensorflow/python/platform/app.py

[loretoparisi]

@girving sorry, didn't get. I'm using the text_classification.py example from the r0.12 blob, where I'm wrong?

[girving]

@loretoparisi You are running text_classification.py from 0.12 against an older version of TensorFlow, which includes an older version of app.py which does not have the flag.

[loretoparisi]

@girving ok this makes sense, in fact my requirements.txt looks like

$ cat requirements.txt 
https://storage.googleapis.com/tensorflow/mac/cpu/tensorflow-0.11.0-py2-none-any.whl
matplotlib
nltk
sklearn
scipy
pandas

Thanks going to move to 0.12 nightly, if there is one?

[jubins]

I am also getting the same error, how can I upgrade my tensorflow?

[loretoparisi]

@jubins for my understanding, the fix is in 0.12 but there is not nightly build in Jenkins already, the latest nightly build is here, but it is still on 0.11, so the only option I see so far is to move to branch r0.12 and compile from the scratch.

[jubins]

@loretoparisi I recompiled my code using r0.12 but it now gives another error 'run() got an unexpected keyword argument argv'

[vrv]

Prior to 1.0, always make sure the tensorflow example code you run is from the same branch as the binary version you have installed.

So your options are:

1. Upgrade your tensorflow installation (via nightly or source built from the same master branch) to use files in the master branch
2. Check out the copy of the example code from the branch you have installed (For example: r0.11 if you have 0.11 installed, r0.10 if you have 0.10 installed).

Mixing and matching source code and binary install will likely cause you pain.

[ttytt1978]

I got it and it works,thanks!

[loretoparisi]

@ttytt1978 how you achieved that?
@vrv it could be useful a step by step guide, thanks.

[ttytt1978]

@loretoparisi

Select the branch match you tensorflow version and download it, run the file in the directory you download.

[marteiro]

Please Dude, read the topic:

git clone https://github.com/tensorflow/tensorflow -b r0.11

[wei-yuan]

Thx a lot every one, it works!

[zcc973784075]

THX all! It helps a lot

[GarethZhang]

@girving Hi, im running tensorflow wheel version 0.10.0 with gpu support (cuda-7.5) and getting the same error. When I tried to branch to r11.0, it successfully installs wheel version 0.11.0 but prompts errors saying that it requires cuda-8.0 libraries. Whereas 0.10.0 only requires cuda-7.5. Do you know if there's another way to fix this error with cuda-7.5 libraries? Thanks!

[zhixinma]

Thanks, i try "-b r0.11" with Git clone and it works

[ghost]

where to clone it

[Zo2m4bie]

Is this problem relevant to version 1.7.0? I installed tensorflow through pycharm and according to print(tf.version) my version is 1.7.0. However I faced the same problem