Multi Label Classification: Customized Pytorch Dataset

In Pytorch, for Multi Label Classification, you need to create a customized Pytorch dataset. This customized Multi Label Classification will be used to iterate over our Multi Label Multi Label Classification data, fetch samples from it, break it into mini-batches, and ultimately, we can use it to create training and test sets for ou Multi Label Classifier (e.g., Artificial Neural Networks)
-------------------------------------------------------------------------------------------------------------------
00:01 Recap and what you will learn
00:23 Intro to the problem of creating customized Pytorch dataset class
00:49 Coding starts
01:02 The source code of the abstract Dataset Class in Pytorch and dunder functions
04:55 Creating the customized dataset
19:18 Over-riding __len__()
20:14 Over-riding __getitem__()
20:14 Creating an object of the customized multi label dataset class
23:22 Understanding the over-ridden __len__() and __getitem__() functions inside the object
25:20 The __add__() dunder function and adding dataset objects
28:18 Outro and what is next
-------------------------------------------------------------------------------------------------------------------
🌎 Website: www.mldawn.com/
🕊 Twitter: / mldawn2018
🔗 Instagram: MLDawn2018
🔗 Linked In: / mehran-ba. .

[@mdbayazid6837]

Thanks for advance coding which is infact rare in RUclips.

[@lenam8048]

I really like your video. Keep going the content ❤

[@SandraBabirye-t7d]

During multilabel classification, how do you handle for the class imbalance

[@MLDawn]

I am no expert in this case but I asked your question from a dear colleague of mine who is an expert in this area. I am sharing his concise and yet informative response to your question:
1. Resampling
2. Different algorithms which handle class imbalance, including cost functions which take care of imbalance
3. Threshold tuning
It's the same as multi class imbalance but more severe.

[@skullgaming5961]

Did you find any solution?

[@pujamane532]

Can you please share the link for above code

[@MLDawn]

My apologies but currently have not dedicated a public repository for this.