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

[DOC] Adds Iñigo's blogposts of Weeks 1 and 2 #39

Merged
merged 4 commits into from
Jun 10, 2024
Merged
Show file tree
Hide file tree
Changes from 2 commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Binary file added _static/images/fibercup_preliminary_results.png
Loading
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
29 changes: 29 additions & 0 deletions posts/2024/2024_06_07_Inigo_week_1.rst
Original file line number Diff line number Diff line change
@@ -0,0 +1,29 @@
First Week into GSoC 2024: Building the AutoEncoder, preliminary results
========================================================================

.. post:: June 7 2024
Copy link
Contributor

Choose a reason for hiding this comment

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

There is still confusion in dates and weeks. If I'm not mistaken there was a blog post (the first one) that corresponded to the community bonding period (week prior to May 27), and since then there have been two weeks: week 1 (ended Fri May 31), and week 2 (ended Fri Jun 7). So I find that this blog post still confuses dates and contents. Can this please be addressed? 2 blog posts are requested. To avoid this in the future, best is to allocate 30-45 minutes every Friday to write about the work that week. I know it is the most scientifically interesting part for the contributor, but it is important for the mentors, to keep track of the project, and the program.

Copy link
Contributor Author

Choose a reason for hiding this comment

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

You are right, I messed up the dates. I will split the info into two blogs. Sorry for the confusion.

:author: Iñigo Tellaetxe
:tags: google
:category: gsoc


What I did this week
Copy link
Member

Choose a reason for hiding this comment

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

Hi @itellaetxe!

Nice blog, clear and concise, Great job.

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Thanks Wachiou!

~~~~~~~~~~~~~~~~~~~~
This week I refactored the AutoEncoder code to match the design patterns and the organization of other Deep Learning models in the DIPY repo. I transferred my code to a `separate repo <https://github.com/itellaetxe/tractoencoder_gsoc>`_ to keep the DIPY repo clean and to experiment freely. Once the final product is working, I will merge it into DIPY. I also packaged the whole repo so I can use it as a library.
Training experiments were run for a maximum of a 150 epochs, with variable results. They are not amazing, but at least we get some reconstruction of the input tracts from FiberCup, which seems to be on the right track. I also implemented training logs that report the parameters I used for training, so I can reproduce the results at any time. This still needs work though, because not all parameters are stored. Need to polish!
The left image shows the input tracts, and the middle and right images show two reconstructions from two different training experiments.

.. image:: /_static/images/fibercup_preliminary_results.png
:alt: Preliminary results of the AutoEncoder training for a subset of plausible fibers of the FiberCup dataset.
:width: 600

What is coming up next week
~~~~~~~~~~~~~~~~~~~~~~~~~~~
With the help of my mentors, we identified possible improvements to the AutoEncoder training process. Yesterday I investigated how PyTorch weights are initialized in convolutional kernels and in Keras Dense layers using the `He Initialization <https://paperswithcode.com/paper/delving-deep-into-rectifiers-surpassing-human>`_. Using custom initializers, one can mimic the same behavior in TensorFlow, which I started to implement also yesterday.
This week should focus on trying to reproduce the small implementation differences that might be causing the model to not converge as the PyTorch one. I will also try to finish implementing the He Initialization in TensorFlow.


Did I get stuck anywhere
~~~~~~~~~~~~~~~~~~~~~~~~
I got a bit stuck refactoring the code to match the DIPY design patterns and also with the TensorFlow implementation itself, because the output shape of the Encoder and the input shape of the Decoder were not matching.
After investigating what caused this issue, I discovered that `tf.shape` was not giving me the usual (and expected) shape of the tensor, conveniently stored in a `tuple`. I found this behavior strange, but I solved the problem just calling a `.shape` method on the Tensor, which does give me the shape tuple I needed.