<div><div dir="auto">yes, i certainly did not mean my message to asad as suggesting we put our full muscle behind python 2.x! until proven wrong, i would like to stick to our bias for 3.x in NLPL—thus, hopefully help encourage folks to transition or start new projects in the modern era.</div></div><div dir="auto"><br></div><div dir="auto">but i was tickled to confirm that i can make one TensorFlow installation (or, technically speaking, an overlay of two installations into the same target directory) work for both python variants. so, the exercise felt instructive :-).</div><div dir="auto"><br></div><div dir="auto">oe</div><div dir="auto"><br></div><div><br><div class="gmail_quote"><div dir="ltr">On Fri, 5 Oct 2018 at 09:50 Martin Matthiesen <<a href="mailto:martin.matthiesen@csc.fi">martin.matthiesen@csc.fi</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div><div style="font-family:arial,helvetica,sans-serif;font-size:10pt;color:#000000"><div>Hello!<br></div><div><br></div><div>Asad is only one user needing 2.7 so far and the code does not seem to be massively forked. I am a bit unsure whether we should officially support Python2.7 with all tools or stick to 3 and support 2.7 on an as-needed basis.<br></div><div>This case was in any case a good reminder that the real world needs sometimes differ from the planning :)<br></div><div><br></div><div>Regards,<br></div><div>Martin<br></div><div><br></div><div>-- <br>Martin Matthiesen<br>CSC - Tieteen tietotekniikan keskus<br>CSC - IT Center for Science<br>PL 405, 02101 Espoo, Finland<br>+358 9 457 2376, <a href="mailto:martin.matthiesen@csc.fi" target="_blank">martin.matthiesen@csc.fi</a><br>Public key : <a href="https://pgp.mit.edu/pks/lookup?op=get&search=0x74B12876FD890704" target="_blank">https://pgp.mit.edu/pks/lookup?op=get&search=0x74B12876FD890704</a><br>Fingerprint: AA25 6F56 5C9A 8B42 009F BA70 74B1 2876 FD89 0704</div><div><br></div><hr id="m_-3148774273241554280zwchr"><div><blockquote style="border-left:2px solid #1010ff;margin-left:5px;padding-left:5px;color:#000;font-weight:normal;font-style:normal;text-decoration:none;font-family:Helvetica,Arial,sans-serif;font-size:12pt"><b>From: </b>"Asad Sayeed" <<a href="mailto:asayeed@mbl.ca" target="_blank">asayeed@mbl.ca</a>><br><b>To: </b>"Stephan Oepen" <<a href="mailto:oe@ifi.uio.no" target="_blank">oe@ifi.uio.no</a>><br><b>Cc: </b>"infrastructure" <<a href="mailto:infrastructure@nlpl.eu" target="_blank">infrastructure@nlpl.eu</a>><br><b>Sent: </b>Thursday, 4 October, 2018 23:40:33<br><b>Subject: </b>Re: [NLPL Task Force (A)] early feedback<br></blockquote></div><div><blockquote style="border-left:2px solid #1010ff;margin-left:5px;padding-left:5px;color:#000;font-weight:normal;font-style:normal;text-decoration:none;font-family:Helvetica,Arial,sans-serif;font-size:12pt"><p>Hi,</p>
<p>Yes, typing python2 or python2.7 works now. Thanks!</p>
<p>Unfortunately I cannot install tensorflow_hub because it thinks
that I am in a virtualenv and I'm pretty sure I am not. But I need
that too. And pyhocon. <br>
</p>
<p>The software I am trying to run is here, if you're wondering:</p>
<p><a class="m_-3148774273241554280moz-txt-link-freetext" href="https://github.com/luheng/lsgn" target="_blank">https://github.com/luheng/lsgn</a><br></p>
<p>It needs the latest stuff. But under python2.7. But it's really
very good...<br>
</p>
<p>Yours,<br>
--Asad.<br>
</p>
<br>
<div class="m_-3148774273241554280moz-cite-prefix">On 2018-10-04 10:27 PM, Stephan Oepen
wrote:<br>
</div>
<blockquote>
<div>
<div dir="auto">i think you are getting the wrong ’python‘
binary; try ’python2‘ or ’python2.7‘. use</div>
</div>
<div dir="auto"><br>
</div>
<div dir="auto">type -all python2</div>
<div dir="auto"><br>
</div>
<div dir="auto">to confirm you end up with the binary from
’.../play/1.11/bin/‘.</div>
<div dir="auto"><br>
</div>
<div dir="auto">oe</div>
<div dir="auto"><br>
</div>
<div><br>
<div class="gmail_quote">
<div dir="ltr">On Thu, 4 Oct 2018 at 22:23 Asad Sayeed <<a href="mailto:asayeed@mbl.ca" target="_blank">asayeed@mbl.ca</a>>
wrote:<br>
</div>
<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Hi,<br>
<br>
I tried it, including the module purge, but then when I
start Python 2.7 <br>
and try to load tensorflow, I get:<br>
<br>
***************<br>
<br>
-bash-4.1$ python<br>
Python 2.7.10 (default, Jul 1 2015, 11:02:23)<br>
[GCC Intel(R) C++ gcc 4.4 mode] on linux2<br>
Type "help", "copyright", "credits" or "license" for more
information.<br>
>>> import tensorflow as tf<br>
Traceback (most recent call last):<br>
File "<stdin>", line 1, in <module><br>
File <br>
"/projects/nlpl/software/play/1.11/lib/python2.7/site-packages/tensorflow/__init__.py",
<br>
line 22, in <module><br>
from tensorflow.python import pywrap_tensorflow #
pylint: <br>
disable=unused-import<br>
File <br>
"/projects/nlpl/software/play/1.11/lib/python2.7/site-packages/tensorflow/python/__init__.py",
<br>
line 49, in <module><br>
from tensorflow.python import pywrap_tensorflow<br>
File <br>
"/projects/nlpl/software/play/1.11/lib/python2.7/site-packages/tensorflow/python/pywrap_tensorflow.py",
<br>
line 74, in <module><br>
raise ImportError(msg)<br>
ImportError: Traceback (most recent call last):<br>
File <br>
"/projects/nlpl/software/play/1.11/lib/python2.7/site-packages/tensorflow/python/pywrap_tensorflow.py",
<br>
line 58, in <module><br>
from tensorflow.python.pywrap_tensorflow_internal
import *<br>
File <br>
"/projects/nlpl/software/play/1.11/lib/python2.7/site-packages/tensorflow/python/pywrap_tensorflow_internal.py",
<br>
line 28, in <module><br>
_pywrap_tensorflow_internal = swig_import_helper()<br>
File <br>
"/projects/nlpl/software/play/1.11/lib/python2.7/site-packages/tensorflow/python/pywrap_tensorflow_internal.py",
<br>
line 24, in swig_import_helper<br>
_mod = imp.load_module('_pywrap_tensorflow_internal',
fp, pathname, <br>
description)<br>
ImportError: <br>
/cluster/software/VERSIONS/python_packages-2.7_6/lib64/libc.so.6: <br>
version `GLIBC_2.14' not found (required by <br>
/projects/nlpl/software/play/1.11/lib/python2.7/site-packages/tensorflow/python/_pywrap_tensorflow_internal.so)<br>
<br>
<br>
Failed to load the native TensorFlow runtime.<br>
<br>
See <br>
<a href="https://www.tensorflow.org/install/install_sources#common_installation_problems" rel="noreferrer" target="_blank">https://www.tensorflow.org/install/install_sources#common_installation_problems</a><br>
<br>
for some common reasons and solutions. Include the entire
stack trace<br>
above this error message when asking for help.<br>
>>><br>
<br>
***************************<br>
<br>
So there is still a glibc discrepancy on the abel login
nodes?<br>
<br>
Thanks!<br>
<br>
Yours,<br>
--Asad.<br>
<br>
On 2018-10-04 01:29 PM, Stephan Oepen wrote:<br>
> hi asad,<br>
><br>
> thanks for your feedback! i am glad you like the Abel
cluster, even<br>
> though it cannot currently run what you would like to
do :-).<br>
><br>
> it seems it did not take very long for an NLPL user to
question our<br>
> assumption that folks should be able to make do with
just Python 3.x<br>
> these days. challenge accepted!<br>
><br>
> i cannot yet promise this will play out in the long
run, nor do i want<br>
> to promise we will commit to supporting it. but it
appears i may have<br>
> managed to apply the ‘glibc gymnastics’ to a TensorFlow
installation<br>
> that supports both Python 2.7 and 3.5.<br>
><br>
> i would encourage you to try the following:<br>
><br>
> module purge<br>
> module use -a /projects/nlpl/software/modulefiles<br>
> module load nlpl-play/1.11/2.7<br>
><br>
> does the above work for you? oe<br>
><br>
> On Thu, Oct 4, 2018 at 12:56 AM Asad Sayeed <<a href="mailto:asayeed@mbl.ca" target="_blank">asayeed@mbl.ca</a>> wrote:<br>
>> Hi,<br>
>><br>
>> I've started trying to use the cluster. It's great
but my current task<br>
>> seems to be impossible there, so I am running it
for longer on local<br>
>> machines. The most recent issue is that I wanted
to run a neural<br>
>> semantic role labeller that requires a recent
tensorflow that runs on<br>
>> python2.7. I was only able to find a tensorflow
that runs on python3,<br>
>> and my attempts to use a singularity container
failed on abel even when<br>
>> I tried to package the container on my local system
with the same<br>
>> singularity version that is on abel. Also
virtualenv and user-level<br>
>> package install did not work for me because of an
incompatible glibc<br>
>> library. Would there be any chance of getting a
tensorflow for python2.7<br>
>> working on abel, along with tensorflow_hub?<br>
>><br>
>> The purpose is to perform SRL on about 10 million
sentences, which<br>
>> should take a day if I can run 400 processes with
10GB memory each (to<br>
>> hold glove vectors). I may in the future have even
larger SRL-related<br>
>> tasks.<br>
>><br>
>> Yours,<br>
>> --Asad.<br>
>><br>
<br>
</blockquote>
</div>
</div>
</blockquote>
<br><br></blockquote></div></div></div></blockquote></div></div>