W1 assignment: tl.Mean for avaraging over vocabulary

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

Hi everyone,

I have a very hard time understanding why is it that tl.Mean should compute an average “over all the words in the vocabulary”, as stated in the W1 Assignment. Could anyone please explain why is it that we do that rather than averaging over each list ?

Thanks a lot (and sry if my question is dumb ! : ) )

2

@ioanaileana You should go through the video of Other Layers in the Week 1 lectures. That explains the how and what to calculate mean for.

3

Thanks @Shantimohan_Elchuri : ) Except that as you can also see in the quiz in the course, the mean is computed over the sentence (tweet in that case) and not the entire vocabulary ! What am I missing?