mratsim / amazon-forest-computer-vision Goto Github PK

I find your code really helpful. Thank you. I see a difference between your kaggle and github code.
In your github code, you declare self.mlb, you put (if I am not wrong) all the total labels in the 'classes' array, However, in your kaggle code, it isn't done and kept empty. What is the difference and and which one of them is correct method?

I try to use this loss ,but get an error as in subject
where is MeanF2loss defined ,
M i missing some thing

Thankyou for your great code!
Recently I am solving a problem about multilabel classify.And i have more lable(500) and is binary and one hot，but the 0 is more than 1, maybe about 50 times?
And i notice that you have use the weight loss and seem that the weight is relevent to the frequency?
Can you give me more details or ideas?
Thankyou best regards!