How to split text data for LSTM training

Question

I\'m using Colab Pro TPU that offers up to 35 GB of memory. My dataset contains 650,000 sequences. I\'m trying to use BiDirectional LSTM to predict the next word.

Whe