Skip to content
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

Environment.yml is broken #110

Closed
wants to merge 2 commits into from
Closed

Conversation

eugenevinitsky
Copy link
Member

Numpy=1.13.0 leads to some crazy looking error. This is extremely high priority, I will confirm that it passes on my machine and @AboudyKreidieh can you try it on your machine and confirm this works?

Numpy=1.13.0 leads to some crazy looking error. This is extremely high priority, I will confirm that it passes on my machine and @AboudyKreidieh can you try it on your machine and confirm this works?
@AboudyKreidieh
Copy link
Member

I got this

UnsatisfiableError: The following specifications were found to be in conflict:
  - matplotlib==2.0.2
  - numpy==1.14.0
  - python==3.5.2
Use "conda info <package>" to see the dependencies for each package.

@AboudyKreidieh
Copy link
Member

do we really need to fix the numpy version? I feel like this is the one package that we can rely on for being stable. open to opinions though

@eugenevinitsky
Copy link
Member Author

The numpy versions can conflict with tensorflow versions.

@eugenevinitsky
Copy link
Member Author

eugenevinitsky commented Sep 6, 2018

We do have to fix it though, if you install the current version set of requirements you immediately get an error.

@eugenevinitsky
Copy link
Member Author

You get that error running which command?

@AboudyKreidieh
Copy link
Member

I see, does tensorflow specify a dependency?

@AboudyKreidieh
Copy link
Member

running conda create env

@eugenevinitsky
Copy link
Member Author

Trying to find it.

@AboudyKreidieh
Copy link
Member

try running the setup instructions on an ubuntu 16.04 machine on aws

@eugenevinitsky
Copy link
Member Author

Ostensibly it's tensorflow==1.10.0 [requires: numpy>=1.13.3,<=1.14.5]
but if you try and import tensorflow with numpy==1.13.3 you get

RuntimeError: module compiled against API version 0xc but this version of numpy is 0xb
ImportError: numpy.core.multiarray failed to import
ImportError: numpy.core.umath failed to import
ImportError: numpy.core.umath failed to import
2018-09-05 23:44:27.419241: F tensorflow/python/lib/core/bfloat16.cc:675] Check failed: PyBfloat16_Type.tp_base != nullptr
[1] 13763 abort python

@eugenevinitsky
Copy link
Member Author

kk, testing ubuntu 16.04

@eugenevinitsky
Copy link
Member Author

Oof yeah you get:

NoPackagesFoundError: Package missing in current linux-64 channels:

  • numpy ==1.14.0
    for conda 4.3.8

@eugenevinitsky
Copy link
Member Author

...how did this ever work?!

@eugenevinitsky
Copy link
Member Author

Confirmed on three machines (@kanaadp @lucfisc and mine) that this works on OSX

This resolves the errors aboudy pointed out. Thanks to @lucfisc for helping with this
@eugenevinitsky
Copy link
Member Author

Confirmed that this works on a 16.04 machine @AboudyKreidieh

@eugenevinitsky
Copy link
Member Author

No luck on a 14.04 machine.

Copy link
Member

@AboudyKreidieh AboudyKreidieh left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM! thanks for doing this

@eugenevinitsky eugenevinitsky deleted the eugenevinitsky-patch-1 branch September 13, 2018 13:21
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

2 participants