Aug 22, 2017 · 1 min read
Hi Rudra,
The code is using pre-trained embeddings provided by the Keras library (they are built-in). If the text contains an unknown word (a word for which there is no embedding), the algorithm drops that word.
I hope I answered your questions!
