Doubt in RNN quiz

#1

You need to train a LSTM Model for Sentiment Analysis. The vocabulary size of your data is given by W1. You have an option using two pre-trained Word Embeddings, with respective vocabulary sizes of W2 and W3. Which Word Embedding would you prefer to use when it is known that W2 << W1 while W3 >> W1

#2

Hey @debjanihome, We will use W3 embeddings, since it will contain feature vector for all words present in W1, It would contain extra words but that word be of no use for us. In W2, many of words from W1 will be missing, so its definitely not the right choice.

HOpe this resolved your doubt.
Plz mark the doubt as resolved in my doubts section. :blush:

#3

okayy so, we’ll use it despite its large unnecessary space consumption?

#4

Hey @debjanihome, yes we will use that.