how to modify the network for multi-classification?

motefly / DeepGBM

SIGKDD'2019: DeepGBM: A Deep Learning Framework Distilled by GBDT for Online Prediction Tasks

647 stars 135 forks source link

how to modify the network for multi-classification? #13

Closed stevenzhang027 closed 4 years ago

stevenzhang027 commented 4 years ago

I'm implementing a version for a multi-classification task, not sure where to change the work. Is it right to change the BatchDense part with parameter out_features from "1" in original codes to n_classes in my case?

motefly commented 4 years ago

In fact, I haven't tried it for multi-classification, because there are #CLASS_NUM-x more trees in LGB for multi-classification, where each part of the trees only contributes to one specifical class. Welcome more explorations in this direction if interest, thank you.

stevenzhang027 commented 4 years ago

already send you an email, looking forward to a discussion with more detail about the direction.

Yang-HangWA commented 4 years ago

So any help for deepGBM on multi-classification task? @stevenzhang027 @motefly