Micro expression and gender classification using Multi-label classification

Hi All, I am applying multi-class classification in micro expression and gender classification. I created my data set with a label like this.

anger_boy, anger_girl, happiness_boy, happiness_girl, repression_girl, repression_boy

However unlike amazon multi-class classification, here only 1 in each class is possible. so is it possible to get the result efficiently in this architecture?

As I am getting 3 labels(boy girl sad) at a time with threshold 0.2 which is not possible in my case.

Please suggest if you have any idea to how to make it better in any way.

Link to the Notebook.

Thanks in Advance.

1 Like