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
Change install to one env file & update docs #484
Conversation
@jcohenadad I had to leave activating the conda env in the installation step, as the installation step needs another command within in ( |
We made the changes to use only the common @vs74, @Stoyan-I-A, @dyt811, @ahill187, @joshuacwnewton, @gisellemartel Could you re-test the installation using this branch? The rendered doc is here. Thank you everyone for your help! |
Installed ADS by referring to the docs of this PR. I had no issues with the installation on MacOS Mojave. Tested |
Installing worked for me on Ubuntu 20.04, as well. 🙂 |
The installation worked. Tested on Windows 10. |
Installation worked for me on MacOS Catalina 10.15.7! Testing the installation however failed for |
Have you used the Theano backend in the past? It looks like the software is trying to use Theano instead of tensorflow. |
Hmm interesting, could this be a tool that was installed with |
I found this, it seems to be an issue when installing Keras through conda-forge, though the post is 3 years old. The only way I found to switch the backend permanently was to explicitly export the variable in |
@gisellemartel could you try this fix temporarily to see if resolves the issue? If so, @mariehbourget let's just make a quick Notice for it on the ReadTheDocs, as it seems like a rare issue and we're migrating away from Keras when we move to IVADOMED soon. |
@gisellemartel, did you get the chance to test the fix describe above? |
Thank you @mariehbourget @mathieuboudreau @Stoyan-I-A for your suggestions. I tried the suggested fix and it worked! I do still get some warnings after running the test however |
Thanks @gisellemartel for the tests. @mathieuboudreau, I added the warning to the doc. Can you review it and let me know if that is ok with you (I can't assign you officially since you opened the PR). Thanks! |
.. NOTE :: For some users, the test may fail because Keras is using Theano backend instead of Tensorflow. In that case, you will see the line ``Using Theano backend.`` when launching ``axondeepseg_test``. To fix this issue, add the line ``export KERAS_BACKEND="tensorflow"`` at the end of the ``<your_conda_install_location>\envs\<your_environment_name>/etc/conda/activate.d/keras_activate.sh`` file, then deactivate and reactivate your environment. The test should print ``Using Tensorflow backend.`` and pass. | ||
|
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Do we need to warn about this?
IIRC, the Keras project has moved on from non-Tensorflow backends ever since it was integrated into TF. They used to have this warning:
Multi-backend Keras has been discontinued. At this time, we recommend that Keras users who use multi-backend Keras with the TensorFlow backend switch to
tf.keras
in TensorFlow 2.0.
So, I would hope that typically Keras users wouldn't run into this issue.
(Though, now that I'm looking into this, they recently edited their README to remove information about this. Now it's "under construction". Hrm.)
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
We decided to warn about this because in particular, one of our test users encountered exactly this issue here in this thread =)
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I guess what I'm asking is, could this be an isolated incident that doesn't apply to most users? 😛
@mariehbourget The message reads well for me; if the tests pass you can approve and merge, which will open the door to merging #479. |
Haha fair. Maybe it coyld be in a troubleshooting section, but lets merge it here for now as the other PR is urgent
Téléchargez Outlook pour Android<https://aka.ms/ghei36>
________________________________
From: Joshua Newton ***@***.***>
Sent: Wednesday, March 24, 2021 3:50:19 PM
To: neuropoly/axondeepseg ***@***.***>
Cc: Mathieu Boudreau ***@***.***>; Mention ***@***.***>
Subject: Re: [neuropoly/axondeepseg] Change install to one env file & update docs (#484)
@joshuacwnewton commented on this pull request.
________________________________
In docs/source/documentation.rst<#484 (comment)>:
+.. NOTE :: For some users, the test may fail because Keras is using Theano backend instead of Tensorflow. In that case, you will see the line ``Using Theano backend.`` when launching ``axondeepseg_test``. To fix this issue, add the line ``export KERAS_BACKEND="tensorflow"`` at the end of the ``<your_conda_install_location>\envs\<your_environment_name>/etc/conda/activate.d/keras_activate.sh`` file, then deactivate and reactivate your environment. The test should print ``Using Tensorflow backend.`` and pass.
+
I guess what I'm saying is, could this be an isolated incident that doesn't apply to most users? 😛
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub<#484 (comment)>, or unsubscribe<https://github.com/notifications/unsubscribe-auth/AAK25ZOZNC2WPMSS4MGIT3TTFIUOHANCNFSM4ZTU5G2A>.
|
Checklist
Resolves #481, Resolves #482, Resolves #483