odlgroup / odlworkshop

Material for ODL workshop Dec 2017

ODL course material

This is the repository for course material (input and output) for the ODL training at KTH in Stockholm, December 2017.

Installing ODL

The latest version of ODL and some supporting libraries are needed in order to use these notebooks. See the installation instructions for further information.

Jupyter Notebooks

Part of the material will be published as Jupyter notebooks. If you're familiar with Mathematica notebooks, you won't be surprised.

To install the notebook software, you can either use conda:

conda install notebook

or pip:

pip install notebook

You can then start the software by running

jupyter notebook

in a terminal. Sometimes one gets OSError: [Errno 99] Cannot assign requested address. In this case an explicit IP has to be given, which should be the equivalent to localhost in your case:

jupyter notebook --ip=127.0.0.1

The notebooks can also be viewed online as static HTML pages. This works (somewhat) on GitHub directly, but better with nbviewer. Just copy the URL of the notebook into the text field at that page.

Feedback

If you find any errors or experience issues, please get back to us (@kohr-h or @adler-j) or create an issue.

odlworkshop's Issues

Broken links to notebooks from maxiv-course-oct-17

The first sentence in "part1_exercises.ipynb" contains two links ("first" and "two"). It looks like both are broken. Probably, they were related to these notebooks, but they are not present in this repo.

Structure the material

We need to move some stuff into folders and add a readme that explains where to start and what all files do.

Make alternative notebooks using pyTorch

Best odl version?

• README.md says "Use the conda variant if possible" (which is 0.5.2 for me on Windows)
• code/part0_install.ipynb says pip install https://github.com/odlgroup/odl/archive/master.zip (which is 0.6.1dev0)

Use rebin package to do binning

https://pypi.python.org/pypi/rebin/1.0.1

Very nice interface

More broken links

Similar to #27 , there are:

• two broken "first notebook" links in part1_using_a_different_solver.ipynb
• a broken "above" link in part1_tv_denoising.ipynb

nomkl & pyfftw on Windows

conda install nomkl and conda install -c conda-forge pyfftw currently fail for me on Windows with PackageNotFoundError. But at least some stuff from the "part1" seem to work. Are they optional actually?

Remove _ syntax

It seems we have

_ = x_true.show('True image')

in quite a few places. I personally prefer

x_true.show('True image');

which does the same thing

Discussion: Should all the notebooks be pre-run online?

I vote yes.

Move repo to odlgroup

Since this counts as "official content" I think it should be on @odlgroup's list of repos.

handout version of lectures

Minor nitpick: it would be a bit easier to read the presentations below if they were recompiled (or have two versions) with "handout" - i.e., using \documentclass[handout]{beamer}

https://github.com/odlgroup/odlworkshop/blob/master/presentations/ODL_intro_HK.pdf

https://github.com/odlgroup/odlworkshop/blob/master/presentations/introduction_JA.pdf

https://github.com/odlgroup/odlworkshop/blob/master/presentations/machine_learning_intro_JA.pdf

Add more dependencies

We should add

• pywavelets
• scikit-learn
• scikit-image
• pytorch

Add learning exercises

We need to add exercises to the deep learning parts.

Installation instructions - trouble

When following the official documentation at

https://github.com/adler-j/odlworkshop/blob/master/code/part0_install.ipynb

(on radon) I get the following error when testing tensorflow (the install proceeded ok.) Possibly radon is out of wack, but no luck with gelfand either (it's out of disk space.)

>>>> import tensorflow
> Traceback (most recent call last):
>  File "/home/kurlberg/anaconda3/lib/python3.6/site-packages/tensorflow/python/pywrap_tensorflow.py", line 58, in <module>
>    from tensorflow.python.pywrap_tensorflow_internal import *
>  File "/home/kurlberg/anaconda3/lib/python3.6/site-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 28, in <module>
>    _pywrap_tensorflow_internal = swig_import_helper()
>  File "/home/kurlberg/anaconda3/lib/python3.6/site-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 24, in swig_import_helper
>    _mod = imp.load_module('_pywrap_tensorflow_internal', fp, pathname, description)
>  File "/home/kurlberg/anaconda3/lib/python3.6/imp.py", line 243, in load_module
>    return load_dynamic(name, filename, file)
>  File "/home/kurlberg/anaconda3/lib/python3.6/imp.py", line 343, in load_dynamic
>    return _load(spec)
> ImportError: /usr/lib/x86_64-linux-gnu/libstdc++.so.6: version `GLIBCXX_3.4.22' not found (required by /home/kurlberg/anaconda3/lib/python3.6/site-packages/tensorflow/python/_pywrap_tensorflow_internal.so)
> 
> During handling of the above exception, another exception occurred:
> 
> Traceback (most recent call last):
>  File "<stdin>", line 1, in <module>
>  File "/home/kurlberg/anaconda3/lib/python3.6/site-packages/tensorflow/__init__.py", line 24, in <module>
>    from tensorflow.python import *
>  File "/home/kurlberg/anaconda3/lib/python3.6/site-packages/tensorflow/python/__init__.py", line 49, in <module>
>    from tensorflow.python import pywrap_tensorflow
>  File "/home/kurlberg/anaconda3/lib/python3.6/site-packages/tensorflow/python/pywrap_tensorflow.py", line 73, in <module>
>    raise ImportError(msg)
> ImportError: Traceback (most recent call last):
>  File "/home/kurlberg/anaconda3/lib/python3.6/site-packages/tensorflow/python/pywrap_tensorflow.py", line 58, in <module>
>    from tensorflow.python.pywrap_tensorflow_internal import *
>  File "/home/kurlberg/anaconda3/lib/python3.6/site-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 28, in <module>
>    _pywrap_tensorflow_internal = swig_import_helper()
>  File "/home/kurlberg/anaconda3/lib/python3.6/site-packages/tensorflow/python/pywrap_tensorflow_internal.py", line 24, in swig_import_helper
>    _mod = imp.load_module('_pywrap_tensorflow_internal', fp, pathname, description)
>  File "/home/kurlberg/anaconda3/lib/python3.6/imp.py", line 243, in load_module
>    return load_dynamic(name, filename, file)
>  File "/home/kurlberg/anaconda3/lib/python3.6/imp.py", line 343, in load_dynamic
>    return _load(spec)
> ImportError: /usr/lib/x86_64-linux-gnu/libstdc++.so.6: version `GLIBCXX_3.4.22' not found (required by /home/kurlberg/anaconda3/lib/python3.6/site-packages/tensorflow/python/_pywrap_tensorflow_internal.so)
> 
> 
> Failed to load the native TensorFlow runtime.
> 
> See https://www.tensorflow.org/install/install_sources#common_installation_problems
> 
> for some common reasons and solutions.  Include the entire stack trace
> above this error message when asking for help.
>>>> 

edit: @adler-j formated