Review and test classification

[ibevers]

This PR is for reviewing, updating, and adding tests to the classification.

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 9.09091% with 40 lines in your changes missing coverage. Please review.

Project coverage is 62.33%. Comparing base (f3c595f) to head (9fab779).
Report is 28 commits behind head on main.

Files	Patch %	Lines
src/tests/audio/tasks/classification_test.py	5.40%	35 Missing ⚠️
...tasks/classification/speech_emotion_recognition.py	28.57%	5 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #106      +/-   ##
==========================================
- Coverage   63.04%   62.33%   -0.72%     
==========================================
  Files          63       65       +2     
  Lines        2073     2156      +83     
==========================================
+ Hits         1307     1344      +37     
- Misses        766      812      +46     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[wilke0818]

I believe this as is might fail with GPU models but can test in colab later. Have been making some changes to these functions in jordan_060 but can definitely merge some of these changes together. also I think batching should possibly be done slightly differently but this is likely a larger conversation with @fabiocat93 next week on how to incorporate best with Pydra. otherwise looks good

[ibevers]

Ah, okay, we'll discuss next week then

[fabiocat93]

@ibevers i see you self-assigned the review of audio classification and started working on it. is there anything else that you think should be done in terms of documentation, tutorial, testing, code optimization?

[ibevers]

@ibevers i see you self-assigned the review of audio classification and started working on it. is there anything else that you think should be done in terms of documentation, tutorial, testing, code optimization?

@fabiocat93 @wilke0818 We still need to discuss the batching approach here. Also, this needs a tutorial.

To-dos:

• Determine whether batching approach makes sense, change if necessary, and test
• Add tutorial

[fabiocat93]

@wilke0818 @ibevers how do you intend to proceed with this? is there any progress update?

[wilke0818]

we never really discussed batching so I don't have any updates. I have a tutorial I can upload at some point for speech emotion recognition

[fabiocat93]

we never really discussed batching so I don't have any updates. I have a tutorial I can upload at some point for speech emotion recognition

would be great. thanks! at least to have it there if someone needs it (even if not optimized)

[fabiocat93]

@satra 's recent questions about speech emotion recognition reminded me that this PR is still open.
@wilke0818 can you please

1. update this code to make it in a mergeable shape
2. refine the speech emotion recognition tutorial (maybe with an example in the categorical space and one in the dimensional space)
3. open an issue describing the batching question (we will consider this as an optimization that we will discuss soon, but not in the context of this PR)

THANK YOU!!

[fabiocat93]

@satra 's recent questions about speech emotion recognition reminded me that this PR is still open. @wilke0818 can you please

1. update this code to make it in a mergeable shape
2. refine the speech emotion recognition tutorial (maybe with an example in the categorical space and one in the dimensional space)
3. open an issue describing the batching question (we will consider this as an optimization that we will discuss soon, but not in the context of this PR)

THANK YOU!!

hi @wilke0818 , any chance you worked on this locally and forgot to push your changes?

[wilke0818]

I mean this code was in mergable shape I think but per #197 will need to be refactored to meet the API specifications of other functionalities. The tutorial is mostly updated but needs a few tweaks based on comments. The batching question is still somewhat open and I'll look back into it either over the holidays or when I get back in January.