Prediction on Causal Language models

Hello, I am trying to make a causal language model in tensorflow. Labels are shifted internally in the model. I need a bit more control over the labels then that. How can I store them somewhere?

1 Like