After the text is converted into integer representation, we’re ready to forward. However, we need to take care of one more thing. Currently, our sentences are of variable length. To feed data into RNN, the length of each input sequence needs to be exactly the same. Let’s see how we can do that.
Now, let’s pad the sentences to a fix length.
You know that a better way (than one-hot vectors) to represent text is word embeddings. Let’s look at how to load word embedding in Python.
All the data preprocessing is now complete. In the next segment, you’ll build an RNN model using the preprocessed data.
Report an error