Exercise 05: log_perplexity

Course Q&A Natural Language Processing NLP with Sequence Models
1

could not understand the purpose of this line:
log_ppx = np.sum(None, None) / np.sum(None, None) # Remember to set the axis properly when summing up

Why do we need a division?

2

@wxlwang , The denominator part is that 1/N in the final step of the formula given just above. Multiplying by 1/N is same as dividing by N, isn’t it?

3

Thanks! I thought it is the mean of all log_probability. Now I got it. It calculates the log_perplexity of each sentence and then mean of corpus’ log_perplexity.