As text is generated, the length gets longer and longer and the computational resources also increase. If it's running on a limited device, it can potentially be a problem. the attention mechanism needs to consider more tokens, which intensifies the computational load.
# Prompt for text generation
sequence = input("start with: ")
# by spaces between words.
window_length = 0 # Set it to 0 in order to consider all the text.