How can we utilize LOW for text classification problem?

cajosantiago / LOW

Official Pytorch implementation of "LOW: Training Deep Neural Networks by Learning Optimal Sample Weights"

8 stars 2 forks source link

How can we utilize LOW for text classification problem? #1

Open bilalbaloch1 opened 2 years ago

bilalbaloch1 commented 2 years ago

I am interested the LOW optimizer for text classification problem. Does your model coordinates the tensorflow/Keras? or we have to utlize this in pytorch only?

cajosantiago commented 2 years ago

LOW can be applied to any loss and in any context. We do not have an official implementation for tensorflow/keras, but it should easy to adapt from our pytorch code.

648374hub commented 10 months ago

Hello, anybody managed to implement the compute weights function to be compatible with tensorflow custom loss class ? Please share. Getting graph execution error while trying.