1

My dataset consists of 5625 Arabic examples and 5625 synsets, and my model is CNN followed by a sigmoid classification layer. I constructed this 5625 synsets to 5625 classes, and my predicted output is a probability between 0 and 1 for each class. What is the loss I should use?

rahma touzi
  • 11
  • 2

1 Answers1

3

Well One Hot Encoder is a great way. What you can do is to create an array of 5625 size with 1 for the synset matching and rest for zero.

Loss here must be binary cross entropy. Its really great for multilabel classification.

Aviral Verma
  • 919
  • 1
  • 4