My query is-
In Emojify_V2, why shouldn’t we use sentences_to_indices() to get the inputs from the sentences we have? Why should we rather use keras.Input()? Then when would be able to send the train dataset to the LSTM to get trained?
Appreciate if someone could provide clarity about it.
Isn’t that what this code does?
(feed in the sentences, outputs the indices, sends them to the embedding layer)
How is the Embedding() layer using these sentence indices, as we are sending only vocab size, embedding dimensions as parameters?
It’s not shown. The assignment uses pretrained embedding weights, so we don’t have to provide the input.
If you have further questions, you may have to post a new thread for another mentor to reply. I’m going to be unavailable for the next few days.