How to set weight value to Crossentropy Loss correctly

For example, I have only two classes, 0 and 1. For class 0, I have 100 samples, for class 1 i have 900 samples. From what I understand, class 0 should have a higher weight, so would the weight [0.9,0.1] be a good approach for training?

I think this question is very similar to yours. Give it a try and feel free to ask any other questions.