Train and Test data sets inconsistency (Exercises: intro to NLP)

[thanujTa]

Hey guys,
I have been trying the challenge in the Intro to NLP chapter which is training our best model with no validation split. Does that mean, we can split test_df into test sentences and labels and use them as validation data in fitting the model? When I did that the error was like this:
ValueError: Data cardinality is ambiguous:
x sizes: 3263
y sizes: 7613
Make sure all arrays contain the same number of samples.

can anyone solve this for me?

thanks in advance.